Re: [Musicpd-dev-team] [PATCH] Add tag support to DSD decoders
On 2012/10/05 09:10, Jurgen Kramer gtmkra...@xs4all.nl wrote: If added these checks to dsdlib_tag_id3(): + /* Prevent broken files causing problems */ + if (is-offset = is-size) + return; + + count = is-size - is-offset; + /* ID3 tag cannot be larger then complete file */ + if ((unsigned)count = is-size) + return; But how does this help? File size is remote input that cannot be trusted, and you don't even need a multi-gigabyte file to get a stack overflow. There is no way to determine how big the tag can be (AFAIK), some tags contain images and are therefore rather big. Large binary tags are not interesting for MPD, they should be skipped. -- Don't let slow site performance ruin your business. Deploy New Relic APM Deploy New Relic app performance management and know exactly what is happening inside your Ruby, Python, PHP, Java, and .NET app Try New Relic at no cost today and get our sweet Data Nerd shirt too! http://p.sf.net/sfu/newrelic-dev2dev ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] [PATCH] Add tag support to DSD decoders
On Sat, 2012-10-06 at 17:27 +0200, Max Kellermann wrote: On 2012/10/05 09:10, Jurgen Kramer gtmkra...@xs4all.nl wrote: If added these checks to dsdlib_tag_id3(): + /* Prevent broken files causing problems */ + if (is-offset = is-size) + return; + + count = is-size - is-offset; + /* ID3 tag cannot be larger then complete file */ + if ((unsigned)count = is-size) + return; But how does this help? File size is remote input that cannot be trusted, and you don't even need a multi-gigabyte file to get a stack overflow. There is no way to determine how big the tag can be (AFAIK), some tags contain images and are therefore rather big. Large binary tags are not interesting for MPD, they should be skipped. OK, I'll artificially limit the max length of the ID3 tag size for both DSD formats as well as limit the max length for the artist and title tags for the DSDIFF format. Average ID3 tag size I encountered is 4034 so I'll set the max to 4096 (maximum I found is 488849). Average tag size I saw for artist and title tags of DSDIFF files is 40 so 60 seems a reasonable limit there. I'll create a new commit with these limits, I'll leave the current file size checks in tact (is-size etc). Jurgen -- Don't let slow site performance ruin your business. Deploy New Relic APM Deploy New Relic app performance management and know exactly what is happening inside your Ruby, Python, PHP, Java, and .NET app Try New Relic at no cost today and get our sweet Data Nerd shirt too! http://p.sf.net/sfu/newrelic-dev2dev ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] [PATCH] Add tag support to DSD decoders
On Thu, 2012-10-04 at 23:02 +0200, Max Kellermann wrote: On 2012/10/04 18:35, Jurgen Kramer gtmkra...@xs4all.nl wrote: I went through the e-mails, I do not recall you having problems with dsdiff_handle_native_tag(). You made the same mistake twice. Maybe I didn't point you to the second instance of the same bug. We discussed dsdlib_tagid3() from dsdlib.c. I've added checks there. I don't see a check that prevents the crash from happening. Where is it? If added these checks to dsdlib_tag_id3(): + /* Prevent broken files causing problems */ + if (is-offset = is-size) + return; + + count = is-size - is-offset; + /* ID3 tag cannot be larger then complete file */ + if ((unsigned)count = is-size) + return; There is no way to determine how big the tag can be (AFAIK), some tags contain images and are therefore rather big. If you have a suggestion please feel free ! For the case in dsdiff_handle_native_tag() here its only for handling artist names and titles and we can probably safely limit the number of characters to a certain number. Jurgen Jurgen -- Don't let slow site performance ruin your business. Deploy New Relic APM Deploy New Relic app performance management and know exactly what is happening inside your Ruby, Python, PHP, Java, and .NET app Try New Relic at no cost today and get our sweet Data Nerd shirt too! http://p.sf.net/sfu/newrelic-dev2dev ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] [PATCH] Add tag support to DSD decoders
On 2012/09/28 09:29, Jurgen Kramer gtmkra...@xs4all.nl wrote: Patch to enable tag support for DSDIFF and DSF DSD decoders. Sent to Max on 19/09. Post here for completeness. Fails to build: src/decoder/dsdlib.c:130:9: error: use of undeclared identifier 'taggoffset'; did you mean 'tagoffset'? -- Don't let slow site performance ruin your business. Deploy New Relic APM Deploy New Relic app performance management and know exactly what is happening inside your Ruby, Python, PHP, Java, and .NET app Try New Relic at no cost today and get our sweet Data Nerd shirt too! http://p.sf.net/sfu/newrelic-dev2dev ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] [PATCH] Add tag support to DSD decoders
On Thu, 2012-10-04 at 09:59 +0200, Max Kellermann wrote: On 2012/09/28 09:29, Jurgen Kramer gtmkra...@xs4all.nl wrote: Patch to enable tag support for DSDIFF and DSF DSD decoders. Sent to Max on 19/09. Post here for completeness. Fails to build: src/decoder/dsdlib.c:130:9: error: use of undeclared identifier 'taggoffset'; did you mean 'tagoffset'? Fixed. I've created a new, clean, commit. Patch to enable tag support for DSDIFF and DSF DSD decoders. http://git.musicpd.org/cgit/gtmkramer/mpd.git/commit/?id=a95346c3eaaa4746f6c0f2f6f62912caed2ad0b7 Jurgen -- Don't let slow site performance ruin your business. Deploy New Relic APM Deploy New Relic app performance management and know exactly what is happening inside your Ruby, Python, PHP, Java, and .NET app Try New Relic at no cost today and get our sweet Data Nerd shirt too! http://p.sf.net/sfu/newrelic-dev2dev ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] [PATCH] Add tag support to DSD decoders
On 2012/10/04 18:13, Jurgen Kramer gtmkra...@xs4all.nl wrote: Fixed. I've created a new, clean, commit. Patch to enable tag support for DSDIFF and DSF DSD decoders. You did not fix the stack overflow bug, did you? I still see it in dsdiff_handle_native_tag(). -- Don't let slow site performance ruin your business. Deploy New Relic APM Deploy New Relic app performance management and know exactly what is happening inside your Ruby, Python, PHP, Java, and .NET app Try New Relic at no cost today and get our sweet Data Nerd shirt too! http://p.sf.net/sfu/newrelic-dev2dev ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] [PATCH] Add tag support to DSD decoders
On 2012/10/04 18:35, Jurgen Kramer gtmkra...@xs4all.nl wrote: I went through the e-mails, I do not recall you having problems with dsdiff_handle_native_tag(). You made the same mistake twice. Maybe I didn't point you to the second instance of the same bug. We discussed dsdlib_tagid3() from dsdlib.c. I've added checks there. I don't see a check that prevents the crash from happening. Where is it? -- Don't let slow site performance ruin your business. Deploy New Relic APM Deploy New Relic app performance management and know exactly what is happening inside your Ruby, Python, PHP, Java, and .NET app Try New Relic at no cost today and get our sweet Data Nerd shirt too! http://p.sf.net/sfu/newrelic-dev2dev ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] [PATCH] Add tag support to DSD decoders - v6
On Mon, 2012-09-03 at 23:09 +0200, Max Kellermann wrote:
On 2012/08/21 20:01, Jurgen Kramer gtmkra...@xs4all.nl wrote:
Can you use dsdlib_tag_id3() for a DoS attack? This looks like it
could easily cause a stack overflow:
+ count = is-size - is-offset;
+ id3_byte_t dsdid3[count];
What is your concern here? The allocation of dsdid3 using count or the
value of count being calculated that way (or both)?
The former. A malicious file can cause a stack overflow. Imagine a
file that makes MPD try to allocate a few gigabytes of heap.
Attached is an old fashioned patch with the updated code against current
mpd git.
I cannot make heads or tales from the mess that is my mpd git repo.
Everything I try only seems to make it worse.
Jurgen
diff -uNrp mpd/src/decoder/dsdiff_decoder_plugin.c mpd-git-0.18-jk/src/decoder/dsdiff_decoder_plugin.c
--- mpd/src/decoder/dsdiff_decoder_plugin.c 2012-09-19 14:30:05.059747095 +0200
+++ mpd-git-0.18-jk/src/decoder/dsdiff_decoder_plugin.c 2012-09-19 14:45:04.204797373 +0200
@@ -52,10 +52,23 @@ struct dsdiff_chunk_header {
uint32_t size_high, size_low;
};
+/** struct for DSDIFF native Artist and Title tags */
+struct dsdiff_native_tag {
+ uint32_t size;
+};
+
struct dsdiff_metadata {
unsigned sample_rate, channels;
bool bitreverse;
uint64_t chunk_size;
+#ifdef HAVE_ID3TAG
+ goffset id3_offset;
+ uint64_t id3_size;
+#endif
+ /** offset for artist tag */
+ goffset diar_offset;
+ /** offset for title tag */
+ goffset diti_offset;
};
static bool lsbitfirst;
@@ -187,6 +200,122 @@ dsdiff_read_prop(struct decoder *decoder
return dsdlib_skip_to(decoder, is, end_offset);
}
+static void
+dsdiff_handle_native_tag(struct input_stream *is,
+ const struct tag_handler *handler,
+ void *handler_ctx, goffset tagoffset,
+ enum tag_type type)
+{
+ if (!dsdlib_skip_to(NULL, is, tagoffset))
+ return;
+
+ struct dsdiff_native_tag metatag;
+
+ if (!dsdlib_read(NULL, is, metatag, sizeof(metatag)))
+ return;
+
+ uint32_t length = GUINT32_FROM_BE(metatag.size);
+ char string[length];
+ char *label;
+ label = string;
+
+ if (!dsdlib_read(NULL, is, label, (size_t)length))
+ return;
+
+ string[length] = '\0';
+ tag_handler_invoke_tag(handler, handler_ctx, type, label);
+ return;
+}
+
+/**
+ * Read and parse additional metadata chunks for tagging purposes. By default
+ * dsdiff files only support equivalents for artist and title but some of the
+ * extract tools add an id3 tag to provide more tags. If such id3 is found
+ * this will be used for tagging otherwise the native tags (if any) will be
+ * used
+ */
+
+static bool
+dsdiff_read_metadata_extra(struct decoder *decoder, struct input_stream *is,
+ struct dsdiff_metadata *metadata,
+ struct dsdiff_chunk_header *chunk_header,
+ const struct tag_handler *handler,
+ void *handler_ctx)
+{
+
+ /* skip from DSD data to next chunk header */
+ if (!dsdlib_skip(decoder, is, metadata-chunk_size))
+ return false;
+ if (!dsdiff_read_chunk_header(decoder, is, chunk_header))
+ return false;
+
+#ifdef HAVE_ID3TAG
+ metadata-id3_size = 0;
+#endif
+
+ /* Now process all the remaining chunk headers in the stream
+ and record their position and size */
+
+ while ( is-offset is-size )
+ {
+ uint64_t chunk_size = dsdiff_chunk_size(chunk_header);
+
+ /* DIIN chunk, is directly followed by other chunks */
+ if (dsdlib_id_equals(chunk_header-id, DIIN))
+ chunk_size = 0;
+
+ /* DIAR chunk - DSDIFF native tag for Artist */
+ if (dsdlib_id_equals(chunk_header-id, DIAR)) {
+ chunk_size = dsdiff_chunk_size(chunk_header);
+ metadata-diar_offset = is-offset;
+ }
+
+ /* DITI chunk - DSDIFF native tag for Title */
+ if (dsdlib_id_equals(chunk_header-id, DITI)) {
+ chunk_size = dsdiff_chunk_size(chunk_header);
+ metadata-diti_offset = is-offset;
+ }
+#ifdef HAVE_ID3TAG
+ /* 'ID3 ' chunk, offspec. Used by sacdextract */
+ if (dsdlib_id_equals(chunk_header-id, ID3 )) {
+ chunk_size = dsdiff_chunk_size(chunk_header);
+ metadata-id3_offset = is-offset;
+ metadata-id3_size = chunk_size;
+ }
+#endif
+ if (chunk_size != 0) {
+ if (!dsdlib_skip(decoder, is, chunk_size))
+break;
+ }
+
+ if ( is-offset is-size ) {
+ if (!dsdiff_read_chunk_header(decoder, is, chunk_header))
+return false;
+ }
+ chunk_size = 0;
+ }
+ /* done processing chunk headers, process tags if any */
+
+#ifdef HAVE_ID3TAG
+ if (metadata-id3_offset != 0)
+ {
+ /* a ID3 tag has preference over the other tags, do not process
+ other tags if we have one */
+ dsdlib_tag_id3(is, handler, handler_ctx, metadata-id3_offset);
+ return true;
+ }
+#endif
+
+ if (metadata-diar_offset != 0)
+ dsdiff_handle_native_tag(is, handler, handler_ctx,
+ metadata-diar_offset, TAG_ARTIST);
+
+ if (metadata-diti_offset != 0)
+ dsdiff_handle_native_tag(is, handler, handler_ctx,
+ metadata-diti_offset, TAG_TITLE);
+ return true;
+}
+
/**
* Read and parse all metadata chunks at the beginning. Stop
Re: [Musicpd-dev-team] [PATCH] Add tag support to DSD decoders - v6
On 2012/09/19 14:50, Jurgen Kramer gtmkra...@xs4all.nl wrote: Attached is an old fashioned patch with the updated code against current mpd git. There's no patch description. I cannot make heads or tales from the mess that is my mpd git repo. Everything I try only seems to make it worse. Please fix your repository. Merging patches manually costs me a lot of time, more than it costs you to learn enough about git to know about its reset command. -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] [PATCH] Add tag support to DSD decoders - v6
On 2012/08/21 20:01, Jurgen Kramer gtmkra...@xs4all.nl wrote: Can you use dsdlib_tag_id3() for a DoS attack? This looks like it could easily cause a stack overflow: + count = is-size - is-offset; + id3_byte_t dsdid3[count]; What is your concern here? The allocation of dsdid3 using count or the value of count being calculated that way (or both)? The former. A malicious file can cause a stack overflow. Imagine a file that makes MPD try to allocate a few gigabytes of heap. -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] [PATCH] Add tag support to DSD decoders - v6
On Mon, 2012-08-20 at 08:44 +0200, Max Kellermann wrote: On 2012/08/16 18:18, Jurgen Kramer gtmkra...@xs4all.nl wrote: Lots of activity in git. Just a reminder for above commit. Memory leak in dsdlib_tag_id3(). Fixed. Added id3_tag_delete Can you use dsdlib_tag_id3() for a DoS attack? This looks like it could easily cause a stack overflow: + count = is-size - is-offset; + id3_byte_t dsdid3[count]; What is your concern here? The allocation of dsdid3 using count or the value of count being calculated that way (or both)? This looks suspicious, too: + uint64_t length = (uint64_t)GUINT32_FROM_BE(metatag.size); + char string[length]; Why is this 32 bit integer casted to 64 bit and then back to 32 bit, anyway? Changed that, length still needs conversion from uint32_t to size_t for usage in dsdlib_read unfortunately. Remember: bad files must *never* cause MPD to crash! If my theory is true, this is a serious security vulnerability you're about to add to MPD. I already have a version laying around where a put in some tighter checks on files being good DSF or DSDIFF files. Maybe I'll sent in a patch for that first. scan_id3_tag() has duplicate documentation. API documentation belongs to the header file. OK, removed comments from .c file Remove the version number from the commit message. I will not merge the old versions of your patch, these should be deleted. Rebase on my master branch, and submit *only* the patches you want me to merge, not all the old versions. I'll remove the version from the patch. Regarding only submitting the needed patches. What the commit I supplied (http://git.musicpd.org/cgit/gtmkramer/mpd.git/commit/?id=9696eff5f075180e88fca504f2502c4f12aef71b) you get more then needed? (I am still not fully proficient with git). Some API documentation for variables like diar_offset, diti_size could be useful. I don't understand what they mean. I'll add comments to the struct. Jurgen -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] [PATCH] Add tag support to DSD decoders - v6
On Tue, Aug 21, 2012 at 08:01:07PM +0200, Jurgen Kramer wrote: Regarding only submitting the needed patches. What the commit I supplied (http://git.musicpd.org/cgit/gtmkramer/mpd.git/commit/?id=9696eff5f075180e88fca504f2502c4f12aef71b) you get more then needed? (I am still not fully proficient with git). Try git rebase -i HEAD~9 on (a copy of) that branch. That will allow you to do several different things with the history in that branch. StGit (stacked git) is another useful tool for managing patchsets in git. (Its usage isn't overly obvious, but it has a good tutorial.) hope that helps, Jonathan Neuschäfer -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] [PATCH] Add tag support to DSD decoders - v6
On 2012/08/16 18:18, Jurgen Kramer gtmkra...@xs4all.nl wrote: Lots of activity in git. Just a reminder for above commit. Memory leak in dsdlib_tag_id3(). Can you use dsdlib_tag_id3() for a DoS attack? This looks like it could easily cause a stack overflow: + count = is-size - is-offset; + id3_byte_t dsdid3[count]; This looks suspicious, too: + uint64_t length = (uint64_t)GUINT32_FROM_BE(metatag.size); + char string[length]; Why is this 32 bit integer casted to 64 bit and then back to 32 bit, anyway? Remember: bad files must *never* cause MPD to crash! If my theory is true, this is a serious security vulnerability you're about to add to MPD. scan_id3_tag() has duplicate documentation. API documentation belongs to the header file. Remove the version number from the commit message. I will not merge the old versions of your patch, these should be deleted. Rebase on my master branch, and submit *only* the patches you want me to merge, not all the old versions. Some API documentation for variables like diar_offset, diti_size could be useful. I don't understand what they mean. -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] [PATCH] Add tag support to DSD decoders
On Thu, 2012-08-09 at 18:29 +0200, Max Kellermann wrote: On 2012/08/09 18:04, Jurgen Kramer gtmkra...@xs4all.nl wrote: Excerpt: In file included from src/input/ffmpeg_input_plugin.c:26:0: /usr/include/ffmpeg/libavformat/avformat.h:158:1: error: ???AVMetadata??? is deprecated (declared at /usr/include/ffmpeg/libavutil/dict.h:41) [-Werror=deprecated-declarations] Hm, bad ffmpeg headers. Which version do you use? I have libav 0.8.3, works fine. OK, I have 0.7.13, I'll probably have to upgrade my system... New attempt: http://git.musicpd.org/cgit/gtmkramer/mpd.git/commit/?id=65c0690891d5312180140bb32c284d973a9fc58e I suggest disabling all of the id3 code when there's no libid3tag, to reduce overhead. For example, the attribute id3_offset should not exist. OK, will do that. I'll revert this commit. The function dsdlib_tag_id3() looks like duplicate code. Better use the scan_id3_tag() function from tag_id3.c. This function is not exported currently. This avoids code duplication, avoids duplicate memory allocation for the temporary tag object, supports all tags, and supports multiple tag values. No problem, I'll look into using scan_id3_tag(). I guess I will have to export it? Jurgen -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] [PATCH] Add tag support to DSD decoders
On Tue, 2012-08-07 at 20:02 +0200, Max Kellermann wrote: On 2012/08/03 15:22, Jurgen Kramer gtmkra...@xs4all.nl wrote: Commit: http://git.musicpd.org/cgit/gtmkramer/mpd.git/commit/?id=0156b6b1717bfc3c43fee6a6a6c6249f2abcb753 Repo: http://git.musicpd.org/cgit/gtmkramer/mpd.git/ Fails to build when libid3tag is disabled: src/decoder/dsdlib.c:142: error: undefined reference to 'id3_tag_parse' src/decoder/dsdlib.c:146: error: undefined reference to 'tag_id3_import' Thanks, fixed it My repo is not in sync with master (won't matter for this patch) how do I keep in sync? Just do a pull of master and push to my git? Please rebase on my master branch (before you begin to work and again right before you submit). Don't create merge commits. I am still not completely comfortable with git, hopefully correct now. Commit: http://git.musicpd.org/cgit/gtmkramer/mpd.git/commit/?id=5cca98871e5ca69cbade555957ad233b79a53aaa Git: http://git.musicpd.org/cgit/gtmkramer/mpd.git/ Jurgen -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] [PATCH] Add tag support to DSD decoders
On 2012/08/08 17:48, Jurgen Kramer gtmkra...@xs4all.nl wrote: On Tue, 2012-08-07 at 20:02 +0200, Max Kellermann wrote: On 2012/08/03 15:22, Jurgen Kramer gtmkra...@xs4all.nl wrote: Commit: http://git.musicpd.org/cgit/gtmkramer/mpd.git/commit/?id=0156b6b1717bfc3c43fee6a6a6c6249f2abcb753 Repo: http://git.musicpd.org/cgit/gtmkramer/mpd.git/ Fails to build when libid3tag is disabled: src/decoder/dsdlib.c:142: error: undefined reference to 'id3_tag_parse' src/decoder/dsdlib.c:146: error: undefined reference to 'tag_id3_import' Thanks, fixed it You added a new commit that is supposed to fix the bug in the first commit, but I do not want to merge the first commit when we already know it's faulty, because it will make bisecting difficult. These two commits should be folded into one before I merge the feature. I am still not completely comfortable with git, hopefully correct now. It's not. You did not rebase, you merged again. Rebasing is as easy as typing: git pull --rebase git://git.musicpd.org/master/mpd.git master -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] [PATCH] Add tag support to DSD decoders
On 2012/08/03 15:22, Jurgen Kramer gtmkra...@xs4all.nl wrote: Commit: http://git.musicpd.org/cgit/gtmkramer/mpd.git/commit/?id=0156b6b1717bfc3c43fee6a6a6c6249f2abcb753 Repo: http://git.musicpd.org/cgit/gtmkramer/mpd.git/ Fails to build when libid3tag is disabled: src/decoder/dsdlib.c:142: error: undefined reference to 'id3_tag_parse' src/decoder/dsdlib.c:146: error: undefined reference to 'tag_id3_import' My repo is not in sync with master (won't matter for this patch) how do I keep in sync? Just do a pull of master and push to my git? Please rebase on my master branch (before you begin to work and again right before you submit). Don't create merge commits. -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] [PATCH] Add tag support to DSD decoders
On Fri, 2012-08-03 at 15:22 +0200, Jurgen Kramer wrote: Commit: http://git.musicpd.org/cgit/gtmkramer/mpd.git/commit/?id=0156b6b1717bfc3c43fee6a6a6c6249f2abcb753 Repo: http://git.musicpd.org/cgit/gtmkramer/mpd.git/ My repo is not in sync with master (won't matter for this patch) how do I keep in sync? Just do a pull of master and push to my git? Managed to sync my repo with master. Latest changes should now be pull-able: http://git.musicpd.org/cgit/gtmkramer/mpd.git/ Commit: http://git.musicpd.org/cgit/gtmkramer/mpd.git/commit/?id=0b576646ba3c6b23c8ba4245153faa0b69132312 Jurgen -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
