Re: [openoffice] 03/03: Fixes from upstream for newer GCC

classic Classic list List threaded Threaded
17 messages Options
Reply | Threaded
Open this post in threaded view
|

Re: [openoffice] 03/03: Fixes from upstream for newer GCC

Matthias Seidel
Hi all,

This commit seems to break building on Windows (with MozillaBuild 3.2).

Including the last lines from output:

cd signtool; /usr/bin/make libs
make[2]: Entering directory
'/cygdrive/c/Source/openoffice/main/nss/wntmsci12.pro/misc/build/nss-3.25/nss/cmd/signtool'
cl -Foout/sign.obj -c -O2 -MD -w44267 -w44244 -w44018 -w44312 -W3
-nologo -D_CRT_SECURE_NO_WARNINGS -D_CRT_NONSTDC_NO_WARNINGS -WX -DXP_PC
-UDEBUG -DNDEBUG -DWIN32 -D_X86_ -D_WINDOWS -DWIN95
-DNSS_NO_INIT_SUPPORT -DUSE_UTIL_DIRECTLY -DNO_NSPR_10_SUPPORT
-DSSL_DISABLE_DEPRECATED_CIPHER_SUITE_NAMES -I../../../dist/out/include
-I../../../dist/public/nss -I../../../dist/private/nss
-I../../../dist/public/seccmd 
"C:/Source/openoffice/main/nss/wntmsci12.pro/misc/build/nss-3.25/nss/cmd/signtool/sign.c"
sign.c
C:/Source/openoffice/main/nss/wntmsci12.pro/misc/build/nss-3.25/nss/cmd/signtool/sign.c(85)
: error C2220: warning treated as error - no 'object' file generated
C:/Source/openoffice/main/nss/wntmsci12.pro/misc/build/nss-3.25/nss/cmd/signtool/sign.c(85)
: warning C4013: 'snprintf' undefined; assuming extern returning int
make[2]: *** [../../coreconf/rules.mk:394: out/sign.obj] Error 2
make[2]: Leaving directory
'/cygdrive/c/Source/openoffice/main/nss/wntmsci12.pro/misc/build/nss-3.25/nss/cmd/signtool'
make[1]: *** [../coreconf/rules.mk:101: libs] Error 2
make[1]: Leaving directory
'/cygdrive/c/Source/openoffice/main/nss/wntmsci12.pro/misc/build/nss-3.25/nss/cmd'
make: *** [coreconf/rules.mk:101: libs] Error 2
dmake:  Error code 2, while making './wntmsci12.pro/misc/build/so_built_nss'

1 module(s):
        nss
need(s) to be rebuilt

Reason(s):

ERROR: error 65280 occurred while making
/cygdrive/c/Source/openoffice/main/nss

When you have fixed the errors in that module you can resume the build
by running:

        build --from nss

Full log can be found here:
https://home.apache.org/~mseidel/nss-error.txt

Regards,

   Matthias


Am 31.10.19 um 14:36 schrieb [hidden email]:

> This is an automated email from the ASF dual-hosted git repository.
>
> arielch pushed a commit to branch AOO42X
> in repository https://gitbox.apache.org/repos/asf/openoffice.git
>
> commit 0803cc9dcc220e6714fbf389f163ad96cd701b38
> Author: Ariel Constenla-Haile <[hidden email]>
> AuthorDate: Thu Oct 31 10:28:47 2019 -0300
>
>     Fixes from upstream for newer GCC
>    
>     Bug 1348767 - logical rather than bitwise OR operator used in OCSP requests
>     Bug 1437734 - sign.c use of sprintf generates format-overflow errors
>     Bug 1438426 - stringop-truncation warning in pathsub.c
>    
>     (cherry picked from commit 4363d9ee5e5c293c9bd85e268df59a8aa7a874c4)
> ---
>  main/nss/makefile.mk           |   5 +-
>  main/nss/nss_bug_1348767.patch |  14 ++++++
>  main/nss/nss_bug_1437734.patch | 107 +++++++++++++++++++++++++++++++++++++++++
>  main/nss/nss_bug_1438426.patch |  12 +++++
>  4 files changed, 137 insertions(+), 1 deletion(-)
>
> diff --git a/main/nss/makefile.mk b/main/nss/makefile.mk
> index c2d51a5..6e717f0 100644
> --- a/main/nss/makefile.mk
> +++ b/main/nss/makefile.mk
> @@ -42,7 +42,10 @@ all:
>  TARFILE_NAME=nss-3.25-with-nspr-4.12
>  TARFILE_MD5=4ec9a36c0f7c9360b149491c013b8d50
>  TARFILE_ROOTDIR=nss-3.25
> -PATCH_FILES=nss.patch
> +PATCH_FILES=nss.patch \
> + nss_bug_1438426.patch \
> + nss_bug_1348767.patch \
> + nss_bug_1437734.patch
>  
>  .IF "$(OS)"=="MACOSX"
>  MACOS_SDK_DIR=$(SDK_PATH)
> diff --git a/main/nss/nss_bug_1348767.patch b/main/nss/nss_bug_1348767.patch
> new file mode 100644
> index 0000000..b776682
> --- /dev/null
> +++ b/main/nss/nss_bug_1348767.patch
> @@ -0,0 +1,14 @@
> +diff -uNrp misc/nss-3.25/nss/lib/libpkix/pkix_pl_nss/pki/pkix_pl_ocsprequest.c misc/build/nss-3.25/nss/lib/libpkix/pkix_pl_nss/pki/pkix_pl_ocsprequest.c
> +--- misc/nss-3.25/nss/lib/libpkix/pkix_pl_nss/pki/pkix_pl_ocsprequest.c 2016-06-20 14:11:28.000000000 -0300
> ++++ misc/build/nss-3.25/nss/lib/libpkix/pkix_pl_nss/pki/pkix_pl_ocsprequest.c 2019-10-27 12:38:20.163600289 -0300
> +@@ -89,8 +89,8 @@ pkix_pl_OcspRequest_Hashcode(
> +         PKIX_HASHCODE(ocspRq->signerCert, &signerHash, plContext,
> +                 PKIX_CERTHASHCODEFAILED);
> +
> +-        *pHashcode = (((((extensionHash << 8) || certHash) << 8) ||
> +-                dateHash) << 8) || signerHash;
> ++        *pHashcode = (((((extensionHash << 8) | certHash) << 8) |
> ++                dateHash) << 8) | signerHash;
> +
> + cleanup:
> +
> diff --git a/main/nss/nss_bug_1437734.patch b/main/nss/nss_bug_1437734.patch
> new file mode 100644
> index 0000000..19e7ead
> --- /dev/null
> +++ b/main/nss/nss_bug_1437734.patch
> @@ -0,0 +1,107 @@
> +--- misc/nss-3.25/nss/cmd/signtool/sign.c 2016-06-20 14:11:28.000000000 -0300
> ++++ misc/build/nss-3.25/nss/cmd/signtool/sign.c 2019-10-28 21:16:32.798336910 -0300
> +@@ -43,6 +43,7 @@ SignArchive(char *tree, char *keyName, c
> +     int status;
> +     char tempfn[FNSIZE], fullfn[FNSIZE];
> +     int keyType = rsaKey;
> ++    int count;
> +
> +     metafile = meta_file;
> +     optimize = _optimize;
> +@@ -81,11 +82,18 @@ SignArchive(char *tree, char *keyName, c
> +         }
> +
> +         /* rsa/dsa to zip */
> +-        sprintf(tempfn, "META-INF/%s.%s", base, (keyType == dsaKey ?
> +-                                                                   "dsa"
> +-                                                                   :
> +-                                                                   "rsa"));
> +-        sprintf(fullfn, "%s/%s", tree, tempfn);
> ++        count = snprintf(tempfn, sizeof(tempfn), "META-INF/%s.%s", base, (keyType == dsaKey ? "dsa" : "rsa"));
> ++        if (count >= sizeof(tempfn)) {
> ++            PR_fprintf(errorFD, "unable to write key metadata\n");
> ++            errorCount++;
> ++            exit(ERRX);
> ++        }
> ++        count = snprintf(fullfn, sizeof(fullfn), "%s/%s", tree, tempfn);
> ++        if (count >= sizeof(fullfn)) {
> ++            PR_fprintf(errorFD, "unable to write key metadata\n");
> ++            errorCount++;
> ++            exit(ERRX);
> ++        }
> +         JzipAdd(fullfn, tempfn, zipfile, compression_level);
> +
> +         /* Loop through all files & subdirectories, add to archive */
> +@@ -95,22 +103,44 @@ SignArchive(char *tree, char *keyName, c
> +     }
> +     /* mf to zip */
> +     strcpy(tempfn, "META-INF/manifest.mf");
> +-    sprintf(fullfn, "%s/%s", tree, tempfn);
> ++    count = snprintf(fullfn, sizeof(fullfn), "%s/%s", tree, tempfn);
> ++    if (count >= sizeof(fullfn)) {
> ++        PR_fprintf(errorFD, "unable to write manifest\n");
> ++        errorCount++;
> ++        exit(ERRX);
> ++    }
> +     JzipAdd(fullfn, tempfn, zipfile, compression_level);
> +
> +     /* sf to zip */
> +-    sprintf(tempfn, "META-INF/%s.sf", base);
> +-    sprintf(fullfn, "%s/%s", tree, tempfn);
> ++    count = snprintf(tempfn, sizeof(tempfn), "META-INF/%s.sf", base);
> ++    if (count >= sizeof(tempfn)) {
> ++        PR_fprintf(errorFD, "unable to write sf metadata\n");
> ++        errorCount++;
> ++        exit(ERRX);
> ++    }
> ++    count = snprintf(fullfn, sizeof(fullfn), "%s/%s", tree, tempfn);
> ++    if (count >= sizeof(fullfn)) {
> ++        PR_fprintf(errorFD, "unable to write sf metadata\n");
> ++        errorCount++;
> ++        exit(ERRX);
> ++    }
> +     JzipAdd(fullfn, tempfn, zipfile, compression_level);
> +
> +     /* Add the rsa/dsa file to the zip archive normally */
> +     if (!xpi_arc) {
> +         /* rsa/dsa to zip */
> +-        sprintf(tempfn, "META-INF/%s.%s", base, (keyType == dsaKey ?
> +-                                                                   "dsa"
> +-                                                                   :
> +-                                                                   "rsa"));
> +-        sprintf(fullfn, "%s/%s", tree, tempfn);
> ++        count = snprintf(tempfn, sizeof(tempfn), "META-INF/%s.%s", base, (keyType == dsaKey ? "dsa" : "rsa"));
> ++        if (count >= sizeof(tempfn)) {
> ++            PR_fprintf(errorFD, "unable to write key metadata\n");
> ++            errorCount++;
> ++            exit(ERRX);
> ++        }
> ++        count = snprintf(fullfn, sizeof(fullfn), "%s/%s", tree, tempfn);
> ++        if (count >= sizeof(fullfn)) {
> ++            PR_fprintf(errorFD, "unable to write key metadata\n");
> ++            errorCount++;
> ++            exit(ERRX);
> ++        }
> +         JzipAdd(fullfn, tempfn, zipfile, compression_level);
> +     }
> +
> +@@ -413,6 +443,7 @@ static int
> + manifesto_xpi_fn(char *relpath, char *basedir, char *reldir, char *filename, void *arg)
> + {
> +     char fullname[FNSIZE];
> ++    int count;
> +
> +     if (verbosity >= 0) {
> +         PR_fprintf(outputFD, "--> %s\n", relpath);
> +@@ -426,7 +457,10 @@ manifesto_xpi_fn(char *relpath, char *ba
> +         if (!PL_HashTableLookup(extensions, ext))
> +             return 0;
> +     }
> +-    sprintf(fullname, "%s/%s", basedir, relpath);
> ++    count = snprintf(fullname, sizeof(fullname), "%s/%s", basedir, relpath);
> ++    if (count >= sizeof(fullname)) {
> ++        return 1;
> ++    }
> +     JzipAdd(fullname, relpath, zipfile, compression_level);
> +
> +     return 0;
> diff --git a/main/nss/nss_bug_1438426.patch b/main/nss/nss_bug_1438426.patch
> new file mode 100644
> index 0000000..978c270
> --- /dev/null
> +++ b/main/nss/nss_bug_1438426.patch
> @@ -0,0 +1,12 @@
> +diff -uNrp misc/nss-3.25/nss/coreconf/nsinstall/pathsub.c misc/build/nss-3.25/nss/coreconf/nsinstall/pathsub.c
> +--- misc/nss-3.25/nss/coreconf/nsinstall/pathsub.c 2016-06-20 14:11:28.000000000 -0300
> ++++ misc/build/nss-3.25/nss/coreconf/nsinstall/pathsub.c 2019-10-27 12:26:03.251950354 -0300
> +@@ -214,7 +214,7 @@ reversepath(char *inpath, char *name, in
> +    xchdir("..");
> + } else {
> +    cp -= 3;
> +-    strncpy(cp, "../", 3);
> ++    memcpy(cp, "../", 3);
> +    xchdir(buf);
> + }
> +     }
>
>


smime.p7s (5K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: [openoffice] 03/03: Fixes from upstream for newer GCC

Ariel Constenla-Haile-3
Hi Matthias,

On Fri, Nov 1, 2019 at 6:17 AM Matthias Seidel
<[hidden email]> wrote:
> C:/Source/openoffice/main/nss/wntmsci12.pro/misc/build/nss-3.25/nss/cmd/signtool/sign.c(85)
> : warning C4013: 'snprintf' undefined; assuming extern returning int

Try with commit 60374adc815ee4dfe39410d357c22db295ee1026
Are the buildbots dead?

Regards
--
Ariel Constenla-Haile
La Plata, Buenos Aires
Argentina

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: [openoffice] 03/03: Fixes from upstream for newer GCC

Matthias Seidel
Hi Ariel,

Am 01.11.19 um 18:20 schrieb Ariel Constenla-Haile:
> Hi Matthias,
>
> On Fri, Nov 1, 2019 at 6:17 AM Matthias Seidel
> <[hidden email]> wrote:
>> C:/Source/openoffice/main/nss/wntmsci12.pro/misc/build/nss-3.25/nss/cmd/signtool/sign.c(85)
>> : warning C4013: 'snprintf' undefined; assuming extern returning int
> Try with commit 60374adc815ee4dfe39410d357c22db295ee1026
Thanks, I will try...
> Are the buildbots dead?

Yes, for some month now...

First the Linux machines were upgraded to Ubuntu 18.04 and stopped
building from SVN.
Windows machines were fine until we switched to Git.

Peter did some changes on the config, but apparently this didn't work out.

Regards,

   Matthias

>
> Regards
> --
> Ariel Constenla-Haile
> La Plata, Buenos Aires
> Argentina
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>


smime.p7s (5K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: [openoffice] 03/03: Fixes from upstream for newer GCC

Matthias Seidel
In reply to this post by Ariel Constenla-Haile-3
Hi Ariel,

Am 01.11.19 um 18:20 schrieb Ariel Constenla-Haile:
> Hi Matthias,
>
> On Fri, Nov 1, 2019 at 6:17 AM Matthias Seidel
> <[hidden email]> wrote:
>> C:/Source/openoffice/main/nss/wntmsci12.pro/misc/build/nss-3.25/nss/cmd/signtool/sign.c(85)
>> : warning C4013: 'snprintf' undefined; assuming extern returning int
> Try with commit 60374adc815ee4dfe39410d357c22db295ee1026

My build finished successfully now, this commit seems to fix it.

Thanks!

Regards,

   Matthias

> Are the buildbots dead?
>
> Regards
> --
> Ariel Constenla-Haile
> La Plata, Buenos Aires
> Argentina
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>


smime.p7s (5K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: [openoffice] 03/03: Fixes from upstream for newer GCC

Don Lewis-2
In reply to this post by Matthias Seidel
On 31 Oct, [hidden email] wrote:

> This is an automated email from the ASF dual-hosted git repository.
>
> arielch pushed a commit to branch AOO42X
> in repository https://gitbox.apache.org/repos/asf/openoffice.git
>
> commit 0803cc9dcc220e6714fbf389f163ad96cd701b38
> Author: Ariel Constenla-Haile <[hidden email]>
> AuthorDate: Thu Oct 31 10:28:47 2019 -0300
>
>     Fixes from upstream for newer GCC
>    
>     Bug 1348767 - logical rather than bitwise OR operator used in OCSP requests
>     Bug 1437734 - sign.c use of sprintf generates format-overflow errors
>     Bug 1438426 - stringop-truncation warning in pathsub.c
>    
>     (cherry picked from commit 4363d9ee5e5c293c9bd85e268df59a8aa7a874c4)
> ---
>  main/nss/makefile.mk           |   5 +-
>  main/nss/nss_bug_1348767.patch |  14 ++++++
>  main/nss/nss_bug_1437734.patch | 107 +++++++++++++++++++++++++++++++++++++++++
>  main/nss/nss_bug_1438426.patch |  12 +++++
>  4 files changed, 137 insertions(+), 1 deletion(-)
>
> diff --git a/main/nss/makefile.mk b/main/nss/makefile.mk
> index c2d51a5..6e717f0 100644
> --- a/main/nss/makefile.mk
> +++ b/main/nss/makefile.mk
> @@ -42,7 +42,10 @@ all:
>  TARFILE_NAME=nss-3.25-with-nspr-4.12
>  TARFILE_MD5=4ec9a36c0f7c9360b149491c013b8d50
>  TARFILE_ROOTDIR=nss-3.25
> -PATCH_FILES=nss.patch
> +PATCH_FILES=nss.patch \
> + nss_bug_1438426.patch \
> + nss_bug_1348767.patch \
> + nss_bug_1437734.patch
>  
>  .IF "$(OS)"=="MACOSX"
>  MACOS_SDK_DIR=$(SDK_PATH)
> diff --git a/main/nss/nss_bug_1348767.patch b/main/nss/nss_bug_1348767.patch
> new file mode 100644
> index 0000000..b776682
> --- /dev/null
> +++ b/main/nss/nss_bug_1348767.patch
> @@ -0,0 +1,14 @@
> +diff -uNrp misc/nss-3.25/nss/lib/libpkix/pkix_pl_nss/pki/pkix_pl_ocsprequest.c misc/build/nss-3.25/nss/lib/libpkix/pkix_pl_nss/pki/pkix_pl_ocsprequest.c
> +--- misc/nss-3.25/nss/lib/libpkix/pkix_pl_nss/pki/pkix_pl_ocsprequest.c 2016-06-20 14:11:28.000000000 -0300
> ++++ misc/build/nss-3.25/nss/lib/libpkix/pkix_pl_nss/pki/pkix_pl_ocsprequest.c 2019-10-27 12:38:20.163600289 -0300
> +@@ -89,8 +89,8 @@ pkix_pl_OcspRequest_Hashcode(
> +         PKIX_HASHCODE(ocspRq->signerCert, &signerHash, plContext,
> +                 PKIX_CERTHASHCODEFAILED);
> +
> +-        *pHashcode = (((((extensionHash << 8) || certHash) << 8) ||
> +-                dateHash) << 8) || signerHash;
> ++        *pHashcode = (((((extensionHash << 8) | certHash) << 8) |
> ++                dateHash) << 8) | signerHash;
> +
> + cleanup:
> +
> diff --git a/main/nss/nss_bug_1437734.patch b/main/nss/nss_bug_1437734.patch
> new file mode 100644
> index 0000000..19e7ead
> --- /dev/null
> +++ b/main/nss/nss_bug_1437734.patch
> @@ -0,0 +1,107 @@
> +--- misc/nss-3.25/nss/cmd/signtool/sign.c 2016-06-20 14:11:28.000000000 -0300
> ++++ misc/build/nss-3.25/nss/cmd/signtool/sign.c 2019-10-28 21:16:32.798336910 -0300
> +@@ -43,6 +43,7 @@ SignArchive(char *tree, char *keyName, c
> +     int status;
> +     char tempfn[FNSIZE], fullfn[FNSIZE];
> +     int keyType = rsaKey;
> ++    int count;
> +
> +     metafile = meta_file;
> +     optimize = _optimize;
> +@@ -81,11 +82,18 @@ SignArchive(char *tree, char *keyName, c
> +         }
> +
> +         /* rsa/dsa to zip */
> +-        sprintf(tempfn, "META-INF/%s.%s", base, (keyType == dsaKey ?
> +-                                                                   "dsa"
> +-                                                                   :
> +-                                                                   "rsa"));
> +-        sprintf(fullfn, "%s/%s", tree, tempfn);
> ++        count = snprintf(tempfn, sizeof(tempfn), "META-INF/%s.%s", base, (keyType == dsaKey ? "dsa" : "rsa"));
> ++        if (count >= sizeof(tempfn)) {
> ++            PR_fprintf(errorFD, "unable to write key metadata\n");
> ++            errorCount++;
> ++            exit(ERRX);
> ++        }
> ++        count = snprintf(fullfn, sizeof(fullfn), "%s/%s", tree, tempfn);
> ++        if (count >= sizeof(fullfn)) {
> ++            PR_fprintf(errorFD, "unable to write key metadata\n");
> ++            errorCount++;
> ++            exit(ERRX);
> ++        }
> +         JzipAdd(fullfn, tempfn, zipfile, compression_level);
> +
> +         /* Loop through all files & subdirectories, add to archive */
> +@@ -95,22 +103,44 @@ SignArchive(char *tree, char *keyName, c
> +     }
> +     /* mf to zip */
> +     strcpy(tempfn, "META-INF/manifest.mf");
> +-    sprintf(fullfn, "%s/%s", tree, tempfn);
> ++    count = snprintf(fullfn, sizeof(fullfn), "%s/%s", tree, tempfn);
> ++    if (count >= sizeof(fullfn)) {
> ++        PR_fprintf(errorFD, "unable to write manifest\n");
> ++        errorCount++;
> ++        exit(ERRX);
> ++    }
> +     JzipAdd(fullfn, tempfn, zipfile, compression_level);
> +
> +     /* sf to zip */
> +-    sprintf(tempfn, "META-INF/%s.sf", base);
> +-    sprintf(fullfn, "%s/%s", tree, tempfn);
> ++    count = snprintf(tempfn, sizeof(tempfn), "META-INF/%s.sf", base);
> ++    if (count >= sizeof(tempfn)) {
> ++        PR_fprintf(errorFD, "unable to write sf metadata\n");
> ++        errorCount++;
> ++        exit(ERRX);
> ++    }
> ++    count = snprintf(fullfn, sizeof(fullfn), "%s/%s", tree, tempfn);
> ++    if (count >= sizeof(fullfn)) {
> ++        PR_fprintf(errorFD, "unable to write sf metadata\n");
> ++        errorCount++;
> ++        exit(ERRX);
> ++    }
> +     JzipAdd(fullfn, tempfn, zipfile, compression_level);
> +
> +     /* Add the rsa/dsa file to the zip archive normally */
> +     if (!xpi_arc) {
> +         /* rsa/dsa to zip */
> +-        sprintf(tempfn, "META-INF/%s.%s", base, (keyType == dsaKey ?
> +-                                                                   "dsa"
> +-                                                                   :
> +-                                                                   "rsa"));
> +-        sprintf(fullfn, "%s/%s", tree, tempfn);
> ++        count = snprintf(tempfn, sizeof(tempfn), "META-INF/%s.%s", base, (keyType == dsaKey ? "dsa" : "rsa"));
> ++        if (count >= sizeof(tempfn)) {
> ++            PR_fprintf(errorFD, "unable to write key metadata\n");
> ++            errorCount++;
> ++            exit(ERRX);
> ++        }
> ++        count = snprintf(fullfn, sizeof(fullfn), "%s/%s", tree, tempfn);
> ++        if (count >= sizeof(fullfn)) {
> ++            PR_fprintf(errorFD, "unable to write key metadata\n");
> ++            errorCount++;
> ++            exit(ERRX);
> ++        }
> +         JzipAdd(fullfn, tempfn, zipfile, compression_level);
> +     }
> +
> +@@ -413,6 +443,7 @@ static int
> + manifesto_xpi_fn(char *relpath, char *basedir, char *reldir, char *filename, void *arg)
> + {
> +     char fullname[FNSIZE];
> ++    int count;
> +
> +     if (verbosity >= 0) {
> +         PR_fprintf(outputFD, "--> %s\n", relpath);
> +@@ -426,7 +457,10 @@ manifesto_xpi_fn(char *relpath, char *ba
> +         if (!PL_HashTableLookup(extensions, ext))
> +             return 0;
> +     }
> +-    sprintf(fullname, "%s/%s", basedir, relpath);
> ++    count = snprintf(fullname, sizeof(fullname), "%s/%s", basedir, relpath);
> ++    if (count >= sizeof(fullname)) {
> ++        return 1;
> ++    }
> +     JzipAdd(fullname, relpath, zipfile, compression_level);
> +
> +     return 0;
> diff --git a/main/nss/nss_bug_1438426.patch b/main/nss/nss_bug_1438426.patch
> new file mode 100644
> index 0000000..978c270
> --- /dev/null
> +++ b/main/nss/nss_bug_1438426.patch
> @@ -0,0 +1,12 @@
> +diff -uNrp misc/nss-3.25/nss/coreconf/nsinstall/pathsub.c misc/build/nss-3.25/nss/coreconf/nsinstall/pathsub.c
> +--- misc/nss-3.25/nss/coreconf/nsinstall/pathsub.c 2016-06-20 14:11:28.000000000 -0300
> ++++ misc/build/nss-3.25/nss/coreconf/nsinstall/pathsub.c 2019-10-27 12:26:03.251950354 -0300
> +@@ -214,7 +214,7 @@ reversepath(char *inpath, char *name, in
> +    xchdir("..");
> + } else {
> +    cp -= 3;
> +-    strncpy(cp, "../", 3);
> ++    memcpy(cp, "../", 3);
> +    xchdir(buf);
> + }
> +     }
>

nss really needs to be updated.  Unfortunately very recent versions
can't be built with some of the ancient compilers that we are stuck with
(like the old version of Visual C++ that we use for 32-bit Windows
builds), or at least not without an insane amount of patching.  I don't
know if gcc on CentOS 6 is also a problem.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: [openoffice] 03/03: Fixes from upstream for newer GCC

Peter Kovacs-3
In reply to this post by Matthias Seidel
I fixed the git statements on Linux. However now perl moduls are
missing. I have asked infra to fix it. Maybe I miss some pointers.


On Windows the issue is that the Git() command checks out the data at a
position where cygwin does not see it.

On the other hand if I pass the command directly no checkout happens.


I am looking for a timeslot to have another round for the windows build
bots.

And to go after the missing modules on Linux.


On 01.11.19 18:37, Matthias Seidel wrote:

> Hi Ariel,
>
> Am 01.11.19 um 18:20 schrieb Ariel Constenla-Haile:
>> Hi Matthias,
>>
>> On Fri, Nov 1, 2019 at 6:17 AM Matthias Seidel
>> <[hidden email]> wrote:
>>> C:/Source/openoffice/main/nss/wntmsci12.pro/misc/build/nss-3.25/nss/cmd/signtool/sign.c(85)
>>> : warning C4013: 'snprintf' undefined; assuming extern returning int
>> Try with commit 60374adc815ee4dfe39410d357c22db295ee1026
> Thanks, I will try...
>> Are the buildbots dead?
> Yes, for some month now...
>
> First the Linux machines were upgraded to Ubuntu 18.04 and stopped
> building from SVN.
> Windows machines were fine until we switched to Git.
>
> Peter did some changes on the config, but apparently this didn't work out.
>
> Regards,
>
>    Matthias
>
>> Regards
>> --
>> Ariel Constenla-Haile
>> La Plata, Buenos Aires
>> Argentina
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: [hidden email]
>> For additional commands, e-mail: [hidden email]
>>

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: [openoffice] 03/03: Fixes from upstream for newer GCC

Ariel Constenla-Haile-3
In reply to this post by Don Lewis-2
Hi Don,

On Fri, Nov 1, 2019 at 8:06 PM Don Lewis <[hidden email]> wrote:
> nss really needs to be updated.  Unfortunately very recent versions
> can't be built with some of the ancient compilers that we are stuck with
> (like the old version of Visual C++ that we use for 32-bit Windows
> builds), or at least not without an insane amount of patching.

Python 2 is going EOL on January 1 2020
https://www.python.org/doc/sunset-python-2/
https://pythonclock.org/

But Python 3 needs a more recent Visual Studio
https://devguide.python.org/setup/#windows
Unless AOO drops support for Python, updating the version of the
Windows compiler seems a priority, python clock is ticking.

> I don't know if gcc on CentOS 6 is also a problem.

It seems Jim was building 4.1.7 on CentOS 5 and 4.2 on 7

Regards
--
Ariel Constenla-Haile
La Plata, Buenos Aires
Argentina

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: [openoffice] 03/03: Fixes from upstream for newer GCC

Ariel Constenla-Haile-3
In reply to this post by Peter Kovacs-3
On Sat, Nov 2, 2019 at 5:53 AM Peter Kovacs <[hidden email]> wrote:
>
> I fixed the git statements on Linux. However now perl moduls are
> missing. I have asked infra to fix it. Maybe I miss some pointers.

Last time I added packages via puppet (*), but things changed since
then, I asked on the INFRA issue. Is there a way to contact infra more
directly?

(*) https://github.com/apache/infrastructure-puppet/commit/66422bb0e21c204a6118e5bba15f028b92fe4e7a

Regards
--
Ariel Constenla-Haile
La Plata, Buenos Aires
Argentina

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: [openoffice] 03/03: Fixes from upstream for newer GCC

Ariel Constenla-Haile-3
In reply to this post by Peter Kovacs-3
On Sat, Nov 2, 2019 at 5:53 AM Peter Kovacs <[hidden email]> wrote:
> On Windows the issue is that the Git() command checks out the data at a
> position where cygwin does not see it.
>
> On the other hand if I pass the command directly no checkout happens.

It's always useful to see the logs:
https://ci.apache.org/builders/openoffice-win7/builds/37/steps/git%20clone%20under%20cygwin/logs/stdio

+ cd /home/buildslave
...
+ git clone --single-branch --branch trunk
https://gitbox.apache.org/repos/asf/openoffice.git build
fatal: destination path 'build' already exists and is not an empty directory.

You have to cd to the right dir, quite strange some line below you
seem to do so...
You also removed the svn revision stuff but left got_revision.
It looks like the underlying OS was changed from Win7 32 bits to Win10
64 bits, that's good news, the buildbot is capable of building
releases.
Quite strange you changed the name (aoo - openoffice) but not the win7 ;)

Now the buildbot stops due to missing NASM
https://ci.apache.org/builders/openoffice-win7/builds/40/steps/configure/logs/stdio
checking nasm.exe assembler path... checking for nasm.exe... no
configure: error: NASM is required to build on Windows, please install
or use --with-nasm-home
program finished with exit code 1
I don't have any machine with Windows to build, so please someone
building master/trunk on Windows (Matthias?) open a jira ticket
requesting to install NASM (once installed, the shell command should
include the new switch).

Regards
--
Ariel Constenla-Haile
La Plata, Buenos Aires
Argentina

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: [openoffice] 03/03: Fixes from upstream for newer GCC

Matthias Seidel
Hi Ariel,

Am 02.11.19 um 18:49 schrieb Ariel Constenla-Haile:

> On Sat, Nov 2, 2019 at 5:53 AM Peter Kovacs <[hidden email]> wrote:
>> On Windows the issue is that the Git() command checks out the data at a
>> position where cygwin does not see it.
>>
>> On the other hand if I pass the command directly no checkout happens.
> It's always useful to see the logs:
> https://ci.apache.org/builders/openoffice-win7/builds/37/steps/git%20clone%20under%20cygwin/logs/stdio
>
> + cd /home/buildslave
> ...
> + git clone --single-branch --branch trunk
> https://gitbox.apache.org/repos/asf/openoffice.git build
> fatal: destination path 'build' already exists and is not an empty directory.
>
> You have to cd to the right dir, quite strange some line below you
> seem to do so...
> You also removed the svn revision stuff but left got_revision.
> It looks like the underlying OS was changed from Win7 32 bits to Win10
> 64 bits, that's good news, the buildbot is capable of building
> releases.
Yes, Windows was updated some years ago, but we didn't change the names
of the builbots... ;-)
> Quite strange you changed the name (aoo - openoffice) but not the win7 ;)
>
> Now the buildbot stops due to missing NASM
> https://ci.apache.org/builders/openoffice-win7/builds/40/steps/configure/logs/stdio
> checking nasm.exe assembler path... checking for nasm.exe... no
> configure: error: NASM is required to build on Windows, please install
> or use --with-nasm-home
> program finished with exit code 1
NASM is a new requirement now. I think it was related to the updated
OpenSSL version.
> I don't have any machine with Windows to build, so please someone
> building master/trunk on Windows (Matthias?) open a jira ticket
> requesting to install NASM (once installed, the shell command should
> include the new switch).
I will do so!

Regards,

   Matthias

>
> Regards


smime.p7s (5K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: [openoffice] 03/03: Fixes from upstream for newer GCC

Don Lewis-2
In reply to this post by Ariel Constenla-Haile-3
On  2 Nov, Ariel Constenla-Haile wrote:

> Hi Don,
>
> On Fri, Nov 1, 2019 at 8:06 PM Don Lewis <[hidden email]> wrote:
>> nss really needs to be updated.  Unfortunately very recent versions
>> can't be built with some of the ancient compilers that we are stuck with
>> (like the old version of Visual C++ that we use for 32-bit Windows
>> builds), or at least not without an insane amount of patching.
>
> Python 2 is going EOL on January 1 2020
> https://www.python.org/doc/sunset-python-2/
> https://pythonclock.org/
>
> But Python 3 needs a more recent Visual Studio
> https://devguide.python.org/setup/#windows
> Unless AOO drops support for Python, updating the version of the
> Windows compiler seems a priority, python clock is ticking.

Yes, we need to upgrade to python 3 for 4.2.x and higher.  This is an
incompatible change, so it's probably not appropriate for 4.1.x.  I've
tried building the FreeBSD port with the system python 3 and the build
fails, so there are probably some other code changes required. I haven't
delved into that at all.

Recent versions of nss are written using c99 features.  It looks like we
need Visual Studio 2015 for that.

What version do we need for python 3?

>> I don't know if gcc on CentOS 6 is also a problem.
>
> It seems Jim was building 4.1.7 on CentOS 5 and 4.2 on 7

I've been building both on CentOS 6.  My CentOS 5 VM broke and I haven't
been able to fix it.

> --
> Ariel Constenla-Haile
> La Plata, Buenos Aires
> Argentina
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: [openoffice] 03/03: Fixes from upstream for newer GCC

Don Lewis-2
In reply to this post by Matthias Seidel
On  2 Nov, Matthias Seidel wrote:

> Hi Ariel,
>
> Am 02.11.19 um 18:49 schrieb Ariel Constenla-Haile:
>> On Sat, Nov 2, 2019 at 5:53 AM Peter Kovacs <[hidden email]> wrote:
>>> On Windows the issue is that the Git() command checks out the data at a
>>> position where cygwin does not see it.
>>>
>>> On the other hand if I pass the command directly no checkout happens.
>> It's always useful to see the logs:
>> https://ci.apache.org/builders/openoffice-win7/builds/37/steps/git%20clone%20under%20cygwin/logs/stdio
>>
>> + cd /home/buildslave
>> ...
>> + git clone --single-branch --branch trunk
>> https://gitbox.apache.org/repos/asf/openoffice.git build
>> fatal: destination path 'build' already exists and is not an empty directory.
>>
>> You have to cd to the right dir, quite strange some line below you
>> seem to do so...
>> You also removed the svn revision stuff but left got_revision.
>> It looks like the underlying OS was changed from Win7 32 bits to Win10
>> 64 bits, that's good news, the buildbot is capable of building
>> releases.
> Yes, Windows was updated some years ago, but we didn't change the names
> of the builbots... ;-)
>> Quite strange you changed the name (aoo - openoffice) but not the win7 ;)
>>
>> Now the buildbot stops due to missing NASM
>> https://ci.apache.org/builders/openoffice-win7/builds/40/steps/configure/logs/stdio
>> checking nasm.exe assembler path... checking for nasm.exe... no
>> configure: error: NASM is required to build on Windows, please install
>> or use --with-nasm-home
>> program finished with exit code 1
> NASM is a new requirement now. I think it was related to the updated
> OpenSSL version.

Yes.

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: [openoffice] 03/03: Fixes from upstream for newer GCC

Bidouille
In reply to this post by Don Lewis-2
Update to Python 3 is reported as an issue since 2014:
https://bz.apache.org/ooo/show_bug.cgi?id=123975

----- Mail original -----

> De: "Don Lewis" <[hidden email]>
> À: [hidden email]
> Envoyé: Samedi 2 Novembre 2019 19:09:17
> Objet: Re: [openoffice] 03/03: Fixes from upstream for newer GCC
>
> On  2 Nov, Ariel Constenla-Haile wrote:
> > Hi Don,
> >
> > On Fri, Nov 1, 2019 at 8:06 PM Don Lewis <[hidden email]>
> > wrote:
> >> nss really needs to be updated.  Unfortunately very recent
> >> versions
> >> can't be built with some of the ancient compilers that we are
> >> stuck with
> >> (like the old version of Visual C++ that we use for 32-bit Windows
> >> builds), or at least not without an insane amount of patching.
> >
> > Python 2 is going EOL on January 1 2020
> > https://www.python.org/doc/sunset-python-2/
> > https://pythonclock.org/
> >
> > But Python 3 needs a more recent Visual Studio
> > https://devguide.python.org/setup/#windows
> > Unless AOO drops support for Python, updating the version of the
> > Windows compiler seems a priority, python clock is ticking.
>
> Yes, we need to upgrade to python 3 for 4.2.x and higher.  This is an
> incompatible change, so it's probably not appropriate for 4.1.x.
>  I've
> tried building the FreeBSD port with the system python 3 and the
> build
> fails, so there are probably some other code changes required. I
> haven't
> delved into that at all.
>
> Recent versions of nss are written using c99 features.  It looks like
> we
> need Visual Studio 2015 for that.
>
> What version do we need for python 3?
>
> >> I don't know if gcc on CentOS 6 is also a problem.
> >
> > It seems Jim was building 4.1.7 on CentOS 5 and 4.2 on 7
>
> I've been building both on CentOS 6.  My CentOS 5 VM broke and I
> haven't
> been able to fix it.
>
> > --
> > Ariel Constenla-Haile
> > La Plata, Buenos Aires
> > Argentina
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: [hidden email]
> > For additional commands, e-mail: [hidden email]
> >
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: [openoffice] 03/03: Fixes from upstream for newer GCC

Matthias Seidel
In reply to this post by Matthias Seidel
Hi Ariel,

Am 02.11.19 um 18:57 schrieb Matthias Seidel:

> Hi Ariel,
>
> Am 02.11.19 um 18:49 schrieb Ariel Constenla-Haile:
>> On Sat, Nov 2, 2019 at 5:53 AM Peter Kovacs <[hidden email]> wrote:
>>> On Windows the issue is that the Git() command checks out the data at a
>>> position where cygwin does not see it.
>>>
>>> On the other hand if I pass the command directly no checkout happens.
>> It's always useful to see the logs:
>> https://ci.apache.org/builders/openoffice-win7/builds/37/steps/git%20clone%20under%20cygwin/logs/stdio
>>
>> + cd /home/buildslave
>> ...
>> + git clone --single-branch --branch trunk
>> https://gitbox.apache.org/repos/asf/openoffice.git build
>> fatal: destination path 'build' already exists and is not an empty directory.
>>
>> You have to cd to the right dir, quite strange some line below you
>> seem to do so...
>> You also removed the svn revision stuff but left got_revision.
>> It looks like the underlying OS was changed from Win7 32 bits to Win10
>> 64 bits, that's good news, the buildbot is capable of building
>> releases.
> Yes, Windows was updated some years ago, but we didn't change the names
> of the builbots... ;-)
>> Quite strange you changed the name (aoo - openoffice) but not the win7 ;)
>>
>> Now the buildbot stops due to missing NASM
>> https://ci.apache.org/builders/openoffice-win7/builds/40/steps/configure/logs/stdio
>> checking nasm.exe assembler path... checking for nasm.exe... no
>> configure: error: NASM is required to build on Windows, please install
>> or use --with-nasm-home
>> program finished with exit code 1
> NASM is a new requirement now. I think it was related to the updated
> OpenSSL version.
>> I don't have any machine with Windows to build, so please someone
>> building master/trunk on Windows (Matthias?) open a jira ticket
>> requesting to install NASM (once installed, the shell command should
>> include the new switch).
> I will do so!
Done:

https://issues.apache.org/jira/browse/INFRA-19373

I have also added you to our Slack channel, I hope you don't mind... ;-)

Matthias

>
> Regards,
>
>    Matthias
>
>> Regards


smime.p7s (5K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: [openoffice] 03/03: Fixes from upstream for newer GCC

Peter Kovacs-3
Slack is the new ASF IRC. Sorry I do not manage to make time. All time I have is eaten by reading mails. :(


Am 2. November 2019 19:54:54 MEZ schrieb Matthias Seidel <[hidden email]>:

>Hi Ariel,
>
>Am 02.11.19 um 18:57 schrieb Matthias Seidel:
>> Hi Ariel,
>>
>> Am 02.11.19 um 18:49 schrieb Ariel Constenla-Haile:
>>> On Sat, Nov 2, 2019 at 5:53 AM Peter Kovacs <[hidden email]>
>wrote:
>>>> On Windows the issue is that the Git() command checks out the data
>at a
>>>> position where cygwin does not see it.
>>>>
>>>> On the other hand if I pass the command directly no checkout
>happens.
>>> It's always useful to see the logs:
>>>
>https://ci.apache.org/builders/openoffice-win7/builds/37/steps/git%20clone%20under%20cygwin/logs/stdio
>>>
>>> + cd /home/buildslave
>>> ...
>>> + git clone --single-branch --branch trunk
>>> https://gitbox.apache.org/repos/asf/openoffice.git build
>>> fatal: destination path 'build' already exists and is not an empty
>directory.
>>>
>>> You have to cd to the right dir, quite strange some line below you
>>> seem to do so...
>>> You also removed the svn revision stuff but left got_revision.
>>> It looks like the underlying OS was changed from Win7 32 bits to
>Win10
>>> 64 bits, that's good news, the buildbot is capable of building
>>> releases.
>> Yes, Windows was updated some years ago, but we didn't change the
>names
>> of the builbots... ;-)
>>> Quite strange you changed the name (aoo - openoffice) but not the
>win7 ;)
>>>
>>> Now the buildbot stops due to missing NASM
>>>
>https://ci.apache.org/builders/openoffice-win7/builds/40/steps/configure/logs/stdio
>>> checking nasm.exe assembler path... checking for nasm.exe... no
>>> configure: error: NASM is required to build on Windows, please
>install
>>> or use --with-nasm-home
>>> program finished with exit code 1
>> NASM is a new requirement now. I think it was related to the updated
>> OpenSSL version.
>>> I don't have any machine with Windows to build, so please someone
>>> building master/trunk on Windows (Matthias?) open a jira ticket
>>> requesting to install NASM (once installed, the shell command should
>>> include the new switch).
>> I will do so!
>
>Done:
>
>https://issues.apache.org/jira/browse/INFRA-19373
>
>I have also added you to our Slack channel, I hope you don't mind...
>;-)
>
>Matthias
>
>>
>> Regards,
>>
>>    Matthias
>>
>>> Regards
Reply | Threaded
Open this post in threaded view
|

Re: [openoffice] 03/03: Fixes from upstream for newer GCC

Ariel Constenla-Haile-3
In reply to this post by Don Lewis-2
Hi Don,

On Sat, Nov 2, 2019 at 3:09 PM Don Lewis <[hidden email]> wrote:
> What version do we need for python 3?

This page https://devguide.python.org/setup/#windows tells VS 2017.
aoo has a lot of external dependencies, I guess that switching to this
version will force updating to newer versions (something not that
trivial, thinking of ICU for example).

Regards
--
Ariel Constenla-Haile
La Plata, Buenos Aires
Argentina

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: [openoffice] 03/03: Fixes from upstream for newer GCC

Don Lewis-2
On  3 Nov, Ariel Constenla-Haile wrote:
> Hi Don,
>
> On Sat, Nov 2, 2019 at 3:09 PM Don Lewis <[hidden email]> wrote:
>> What version do we need for python 3?
>
> This page https://devguide.python.org/setup/#windows tells VS 2017.
> aoo has a lot of external dependencies, I guess that switching to this
> version will force updating to newer versions (something not that
> trivial, thinking of ICU for example

Upgrading icu is something else that we really need to do.  I looked at
it once a long time ago and it is definitely not trivial.  Probably
beyond my skill set.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]