Activity for Andrew J. Schorr
-
Andrew J. Schorr committed [04bfd1] on Code
New release 1.1.2 of xml to fix a bug in xml-conv-enc's logic to find an
-
Patch lmdb and select rules for generating header files to use $(GAWKPROG) instead of $(AWK) for more predictable behavior and better portability.
-
gawk-lmdb: basic test fails on macOS
-
Thanks for the bug report. I patched this test case, and the new release 1.1.3 should work on MacOS.
-
New release 1.1.3 of lmdb to fix MacOS compatibility issues.
-
Patch lmdb to find lmdb.h in nonstandard locations and to get the test case to work on MacOS.
-
gawk-select: outputfd and nonblock tests fail on macOS
-
I got access to a MacOS system on the GCC compile farm. I am unable to depulicate the "nonblock" error. As best I can tell, that must have been caused by an extremely sluggish system. Are you able to replicate the problem? I patched the "outputfd" test case to use "cksum" instead of "wc -l", since "cksum" seems to have a more consistent output format. Please try gawk-select-1.1.4. Regards, Andy
-
New release 1.1.4 of select to fix MacOS compatibility issue.
-
Fix a couple of minor portability issues in errno and select to improve MacOS support.
-
New release 1.0.1 of reclen to roll up a bunch of minor changes.
-
Hi, sorry for the slow response. As you say, this is most likely just a minor platform difference in error messages. Is this still a concern? I'm guessing the extension works OK. If it's causing trouble, we can try to get it fixed. Regards, Andy
-
New release 1.1.3 of select to improve Makefile portability and robustness.
-
Sorry for the long delay. The first "nonblock" error looks like a weird timing issue. Is the system extremely sluggish? The second issue seems to be caused by unexpected output formatting from "wc -l". In both cases, I think it's not important, and the extension is in fact working properly. Is this still a concern? If so, we can try to fix these nits, but it's probably not important. Regards, Andy
-
Update json spec file to require gawk >= 5.0.0 because it uses namespaces.
-
New release 1.1.1 of errno to improve Makefile robustness.
-
gawk-lmdb: failure to parse headers on macOS
-
I think this should be fixed in lmdb-1.1.2. Please test that it solves the issue. Thanks, Andy
-
lmdb 1.1.1: test suite fails
-
I uploaded a new release lmdb-1.1.2 that should fix the problem (a filename collision when running the tests in parallel).
-
New release 1.1.2 of lmdb to improve Makefile robustness and fix a problem with running the test suite in parallel.
-
Commit some old changes to shared/gawkext.m4 that improve configure help messages and add some new options to make it easier to port to MinGW.
-
Actually, what I'm finding is that it fails frequently when I run "make -j check", but not single-threaded "make check". I think this is most likely because dict.awk and dict_cursor.awk are both using the same filename "./test.lmdb". Assuming this is the issue, I will patch and upload a new release. Sorry for the brain damage. Regards, Andy On Tue, Jan 16, 2024 at 03:38:05PM -0000, Andrew J. Schorr wrote: I upgraded to Fedora 39 to see if the issue persists. What I'm finding is that it's intermittent....
-
I upgraded to Fedora 39 to see if the issue persists. What I'm finding is that it's intermittent. Sometimes "make check" passes, and other times it doesn't. Is it also intermittent on your platform? Regards, Andy On Tue, Jan 16, 2024 at 02:11:41PM -0000, Andrew J. Schorr wrote: Hi, It works OK for me on CentOS 7 and on CentOS Stream 9. What specific version of Linux are you using? And what version of gawk, gawkextlib, and lmdb are you using? What's the exact command you are running? If I understand...
-
Hi, It works OK for me on CentOS 7 and on CentOS Stream 9. What specific version of Linux are you using? And what version of gawk, gawkextlib, and lmdb are you using? What's the exact command you are running? If I understand more about your environment, that may give me a clue... That being said, I just tested on Fedora 38, and it seems to be broken there. So I will try to dig deeper when I get some time. Regards, Andy On Tue, Jan 16, 2024 at 08:31:21AM -0000, Tomasz Kłoczko wrote: I can't duplicate...
-
Fix pgsql test to work with postgresql 11, and bump pgsql version to 1.1.2.
-
Hi, I can't duplicate the problem on my system. What platform are you using? Did you run this as part of rpmbuild? If so, please provide full details of what you did so I can try to recreate the issue. What version of lmdb is installed on your system? And what version of gawk and gawkextlib? Regards, Andy On Mon, Jan 15, 2024 at 10:15:38AM -0000, Tomasz Kłoczko wrote: ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ [bugs:#10] lmdb 1.1.1: test suite fails Status: open...
-
Delete obsolete timex extension, since the functions are now included in the core time extension.
-
Update make_extension_directory.sh and */packaging/gawk*.spec.in to reflect that the gawk_api_major_version was bumped from 3 to 4.
-
gawk-json-2.0.0: from_json fails parsing simple multidimensional array
-
Thanks for the bug report. Arnold patched it. The bug should be fixed in gawk-json-2.0.1, which I just uploaded. Regards, Andy
-
Thanks for the bug report. I see the same issue on Linux using the gawk master branch: Program received signal SIGSEGV, Segmentation fault. make_str_node (s=0x0, len=0, flags=<optimized out="">) at node.c:422 422 r->stptr[len] = '\0'; (gdb) bt</optimized> 0 make_str_node (s=0x0, len=0, flags=<optimized out="">) at node.c:422</optimized> 1 0x000000000044fc7c in api_set_array_element (id=<optimized out="">, value=0x7fffffffd1b0, index=0x7fffffffd190,</optimized> a_cookie=0x4e3990) at gawkapi.c:1106...
-
Issue new release 1.0.3 of the gd library incorporating the
-
gawk-gd translucency fix
-
I'll push out a new release as well.
-
Thanks. The patch has been committed.
-
Fix bug in gd/gd.c:do_gdImagePngName to add call to gdImageSaveAlpha.
-
Hi, Thanks for your report. But isn't "fts" part of the filefuncs library that's packaged with gawk? Issues with gawk extensions should be reported to help-gawk@gnu.org or by using the gawkbug script. Please see: https://www.gnu.org/software/gawk/ In any case, it would be preferable if you can include a short sample program demonstrating the problem you are encountering. Best, Andy On Sat, Jun 17, 2023 at 04:34:35PM -0000, Evgeny Dobromilsky wrote: There is a bug or issue. When I use such 2 paths...
-
Hi Jürgen, I did notice his unusual locale settings. I configured my locale to match his settings, but it did not change the result -- the script works fine for me. So I don't know what's causing his problem. Regards, Andy On Wed, Jun 14, 2023 at 10:31:45PM -0000, "Jürgen Kahrs" wrote: Hello Andrew, did you notice the unusual locale ? Mostly German settings, but some U.S. settings. LANG=en_US.UTF-8 LANGUAGE= LC_CTYPE="en_US.UTF-8" LC_COLLATE="en_US.UTF-8" Hard to imagine that this combination causes...
-
Hi, I'm unable to duplicate your problem. I downloaded gawk-5.2.2 and gawk-json-2.0.0, and it all just works fine for me. Valgrind does not find any errors either. I also tested against the gawk master branch, and still no problems. Here are the steps I followed: 1. Download gawk-5.2.2 and json-2.0.0 2. cd /extra_disk/tmp 3. tar xf ~/Downloads/gawk-5.2.2.tar.xz 4. cd gawk-5.2.2 5. ./configure --prefix=/extra_disk/tmp/gawk522 && make -j && make -j check && make -j install && echo SUCCESS 6. cd .....
-
OK, duh, I see that it's a SEGV in the subject, so you can ignore my first question.
-
Hi, What error message are you getting? And what versions of gawk and the 2 libraries are you using? I cannot reproduce the issue because I can't even get past the first line where it sets the value of json. I'm getting an error about invalid char in expression. Can you please post your reproducing program as an attachment? I guess maybe it's getting garbled at some point, and/or there are locale issues. What locale are you using? Regards, Andy
-
Patch shared/test.makefile to work on MinGW
-
The attached solution avoids set -o pipefail, so should be more robust, at the expense of introducing an intermediate temporary file, which is a bit inelegant. Does this solution work well for you?
-
I don't have access to MacOS. On Linux, the attached patch fixes the problem with not detecting the error. Does it work on MacOS? I don't know how portable the "set -o pipefail" bash feature is.
-
Thanks for the bug report. It's also in the errno extension. I will fix all 3, but operating a bit slowly now due to a biking accident that resulted in a concussion. Regards, Andy
-
Patch select library to parse non-numeric timeout args properly as infinite timeouts, and release new version 1.1.2.
-
Update make_extension_directory.sh and */packaging/gawk*.spec.in to reflect that the gawk_api_major_version was bumped from 2 to 3.
-
json: commit some final tweaks for version 2.0.0 using namespaces.
-
Thanks. In the future, in your git commit comments, can you please mention the name of the extension that's being modified?
-
Bug fix, regex cache and other enhancements
-
On Sat, Sep 12, 2020 at 07:05:59AM -0000, Jannick wrote: First of all, sorry for the late reply. No worries. I'm sure we're all pretty busy with other issues. Yes - the tag gawk-xml-1.1.1 is exactly what I meant. Thanks! You're welcome. git provides ample features to create these tags within a couple of lines, so not a real burden as far as I can tell. It's not a big burden. It's just a question of remembering to do it. Instead of putting the version number into the snapshot name (which would be...
-
reclen: remove unused variable fixed_buffer in reclen_get_record.
-
I have no particular objection to attempting to add tags for releases, but I don't see why it particularly matters, in the sense that there are almost no downloads of any of this stuff anyway. Is anybody actually using the snapshot releases? The stats show zero downloads for those. And frankly, there are almost no downloads of any of the released extensions either. So it seems like we're talking about adding more administrative burden for a project that almost nobody uses. Jürgen created the snapshot...
-
Hi Jannick, Thanks for the suggestion, but please help me to understand better the problem you're trying to solve. Currently, when there's a new release of a package, the procedure is to update the version number in configure.ac in the AC_INIT and make an entry in the NEWS file. If I understand correctly, I think you're asking us to add a git tag as well labeling the release. Which problem does that solve? Are you suggesting that we remove the explicit version number in configure.ac:AC_INIT and instead...
-
Patch timex man page to clarify strptime's return value.
-
Add new reclen extension from Arnold.
-
timex extension: check whether system has strptime, fix test that breaks on systems with timezone idiocy, and define _POSIX_C_SOURCE instead of __USE_POSIX199309
-
Remove obsolete strptime extension now subsumed by timex.
-
Patch timex/timex.c to eliminate a few gcc warnings.
-
timex: remove derived files that are auto-generated.
-
Thanks. I merged it. In the future, please update the NEWS file when you issue a new release with a short message explaining the major changes.
-
Fixed memory leak
-
Fix mpfr spec file.
-
Upgrade mpfr package to API version 2
-
Added aregex extension files
-
Add missing file mbs/m4/gawkext.m4.
-
Redis extension, new release
-
Merge commit '059fa97898d31aabcd42b4bafe7bbd9a65282acf'
-
Upgrade xml to 1.1.1 to fix some packaging issues.
-
Upgrade pgsql to 1.1.1 to fix some packaging issues.
-
Release lmdb version 1.1.1 for some minor packaging updates and man page fixes.
-
Patch top-level dirloop.sh script to ignore new _web subdirectory.
-
Redis extension, bumping.
-
Update json to version 1.0.2 now that gawk 4.2.1 has been released and we no longer need a private copy of gawkapi.h.
-
Merge commit '20cd07af82774592d68ff066da67db2a38c0f71e'
-
Update all packaging spec files to add BuildRequires: gcc or gcc-c++, since Fedora 29 will require this to be explicit.
-
I don't think this really makes sense in the context of this project, since each subdirectory has its own releases.
-
No release tags in git repository
-
No longer relevant.
-
Should xgawk try to install in the same place as gawk?
-
Patch make_extension_directory.sh to enhance rpm spec file generation.
-
Wrap line in abort specfile to avoid rpmlint complaint.
-
Upgrade abort spec file to Fedora standards.
-
Fix redis packaging to remove unnecessary dependency on gawkextlib
-
New release of abort extension with man page improvement and build fix.
-
Patch abort to avoid the need for gawkextlib.h, and patch make_extension_directory.sh to do this properly going forward.
-
Minor cleanups for abort extension.
-
Replace incorrect 'awk_ext_id_t *' with 'awk_ext_id_t' in several places.
-
Upgrade redis extension to API version 2
-
Update lmdb and pgsql rpm spec files to stop attempting to install translations
-
Remove nonexistent French translations from csv, gd, haru, lmdb, mpfr, pgsql, and redis.
-
New nl_langinfo release 1.1.0 for Fedora inclusion.
-
Fix select test case that could fail on extremely slow machines.
-
In select packaging, add a couple more examples as doc files.
-
Upgrade select extension to API version 2
-
Minor tweaks to errno extension to eliminate dependency on gawkextlib.h
-
Fix errno packaging bug
-
Minor improvement to errno man page
-
Upgrade errno package to API version 2
1 >
