Message ID | 20240409000546.3628898-1-jacob.e.keller@intel.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | mailsplit add option to include sanitized subject in filename | expand |
Jacob Keller <jacob.e.keller@intel.com> writes: > I originally wanted to avoid the need for an option, but git-am > currently depends on the strict sequence number filenames. It is > unclear how difficult it would be to refactor git-am to work with > names that include the extra subject data. I am not sure if I follow. Do you mean $ git am ./dir/0*.txt in a directory where I already have these files $ ls dir/0*.txt dir/0001-Documentation-CodingGuidelines.txt dir/0002-quote-assigned-value.txt dir/0003-t-local-var.txt that have one patch per message does not work?
On Mon, Apr 8, 2024 at 6:55 PM Junio C Hamano <gitster@pobox.com> wrote: > > Jacob Keller <jacob.e.keller@intel.com> writes: > > > I originally wanted to avoid the need for an option, but git-am > > currently depends on the strict sequence number filenames. It is > > unclear how difficult it would be to refactor git-am to work with > > names that include the extra subject data. > > I am not sure if I follow. Do you mean > > $ git am ./dir/0*.txt > No, I mean git am invokes git mailsplit to split a mailbox file into a temporary directory, and then expects to find exactly "0000", "0001", "0002" etc, but not "0001-fix-bug" and "0002-implement-feature"
Jacob Keller <jacob.keller@gmail.com> writes: > No, I mean git am invokes git mailsplit to split a mailbox file into a > temporary directory, and then expects to find exactly "0000", "0001", > "0002" etc, but not "0001-fix-bug" and "0002-implement-feature" Ah, of course. "am" invokes mailsplit with the understanding that its external interface is that it will get the total number as decimal number from its standard output, and the files are named as just numbers in the specified directory, with specified precision. If you are mucking with mailsplit to update its output, of course you must update the expected way "am" receives its input.
Jacob Keller <jacob.keller@gmail.com> writes: >> > I originally wanted to avoid the need for an option, but git-am >> > currently depends on the strict sequence number filenames. It is >> > unclear how difficult it would be to refactor git-am to work with >> > names that include the extra subject data. The change may be a bit involved but depending on where you decide to stop, it may not be too bad. What is your design goal of this topic? IOW, what is the maximum corrupted ordering of patches in a single mailbox do you want to recover from? The easiest and cleanest would be if you assume that the messages are in scrambled order, but are all from the same series, correctly numbered, without anything missing. A mbox may have 8 patches from a 8-patch series, with their subject lines having [1/8] to [8/8] without duplicates or droppages, without any other message that does not belong to the series. If that is where you are willing to stop, then you can still name the individual messages with just numbers (but taken out of the subject line, not the order the input was splitted into). "am" does not have to even know or care what you are doing in mailsplit in this case. THe next level would be to still assume that you stop at the same place (i.e. you do not support patches from multiple series in the same mailbox), but use the number-santized-subject format. This would be a bit more involved, but I think all you need to update on the "am" side is where the am_run() assigns the message file to the local variable "mail". You know the temporary directory where you told "mailsplit" to create these individual messages, so you should be able to "opendir/readdir/closedir" and create a list of numbered files in the directory very early in "git am". Knowing msgnum(state) at that point in the loop, it should be trivial to change the code that currently assumes the 4-th file is named "0004" to check for the file whose name begins with "0004-". I personally am not at all interested in doing that myself, because I do not see a reasonable way to lift the limitation of allowing a mailbox holding patches from only one series, and if we assume that a tool (i.e. "am" driving "mailsplit" in the new mode) with such a limitation is still useful, the source of such a scrambled mailbox must be quite a narrow and common one. At that point, I suspect that fixing the scrambling at that narrow and common source (e.g. your "t.mbox.gz from public inbox server that cannot be told to sort the messages in any order other than the arrival timestamp") would be a much better use of our engineering resource.
On 4/10/2024 10:55 PM, Junio C Hamano wrote: > Jacob Keller <jacob.keller@gmail.com> writes: > >>>> I originally wanted to avoid the need for an option, but git-am >>>> currently depends on the strict sequence number filenames. It is >>>> unclear how difficult it would be to refactor git-am to work with >>>> names that include the extra subject data. > > The change may be a bit involved but depending on where you decide > to stop, it may not be too bad. > > What is your design goal of this topic? IOW, what is the maximum > corrupted ordering of patches in a single mailbox do you want to > recover from? > > The easiest and cleanest would be if you assume that the messages > are in scrambled order, but are all from the same series, correctly > numbered, without anything missing. A mbox may have 8 patches from > a 8-patch series, with their subject lines having [1/8] to [8/8] > without duplicates or droppages, without any other message that does > not belong to the series. If that is where you are willing to stop, > then you can still name the individual messages with just numbers > (but taken out of the subject line, not the order the input was > splitted into). "am" does not have to even know or care what you > are doing in mailsplit in this case. This is the main problem I'd like to solve. My original proposal was to try and do just this, but the logic for extracting the number was bad. Maybe just directly using the subject-based name and sorting by that using standard alpha-numeric sort would be sufficient? > > THe next level would be to still assume that you stop at the same > place (i.e. you do not support patches from multiple series in the > same mailbox), but use the number-santized-subject format. This > would be a bit more involved, but I think all you need to update on > the "am" side is where the am_run() assigns the message file to the > local variable "mail". You know the temporary directory where you > told "mailsplit" to create these individual messages, so you should > be able to "opendir/readdir/closedir" and create a list of numbered > files in the directory very early in "git am". Knowing msgnum(state) > at that point in the loop, it should be trivial to change the code > that currently assumes the 4-th file is named "0004" to check for > the file whose name begins with "0004-". Yea, we pretty much just have to get the git-am process to work with the new names. I can look at using opendir/readdir here instead. > > I personally am not at all interested in doing that myself, because > I do not see a reasonable way to lift the limitation of allowing a > mailbox holding patches from only one series, and if we assume that > a tool (i.e. "am" driving "mailsplit" in the new mode) with such a > limitation is still useful, the source of such a scrambled mailbox > must be quite a narrow and common one. At that point, I suspect > that fixing the scrambling at that narrow and common source (e.g. > your "t.mbox.gz from public inbox server that cannot be told to sort > the messages in any order other than the arrival timestamp") would > be a much better use of our engineering resource. > Ya I don't care much about multiple series. I care more about making it handle scrambled series better than it does now. I download series off of lore.kernel.org (public-inbox based) and those seem to routinely have series out-of-order. I suspect this is because it bases them on arrival date and sometimes certain mailers get it out of order when sending.
Jacob Keller <jacob.e.keller@intel.com> writes: >> THe next level would be to still assume that you stop at the same >> place (i.e. you do not support patches from multiple series in the >> same mailbox), but use the number-santized-subject format. This >> would be a bit more involved, but I think all you need to update on >> the "am" side is where the am_run() assigns the message file to the >> local variable "mail". You know the temporary directory where you >> told "mailsplit" to create these individual messages, so you should >> be able to "opendir/readdir/closedir" and create a list of numbered >> files in the directory very early in "git am". Knowing msgnum(state) >> at that point in the loop, it should be trivial to change the code >> that currently assumes the 4-th file is named "0004" to check for >> the file whose name begins with "0004-". > > Yea, we pretty much just have to get the git-am process to work with the > new names. I can look at using opendir/readdir here instead. Not "here", but probably just after you called "mailsplit" and saw it return. After that nobody should be adding more split mail messages to the directory, so you do it once to grab all filenames. > Ya I don't care much about multiple series. I care more about making it > handle scrambled series better than it does now. I download series off > of lore.kernel.org (public-inbox based) and those seem to routinely have > series out-of-order. I suspect this is because it bases them on arrival > date and sometimes certain mailers get it out of order when sending. Yeah, and that is why I said it would be a better use of the engineering resource to fix it at the source. Such a fix will benefit folks with existing versions of "git am", not needing to wait for your improved version. Thanks.
On Thu, Apr 11, 2024 at 2:25 PM Junio C Hamano <gitster@pobox.com> wrote: > > Jacob Keller <jacob.e.keller@intel.com> writes: > > >> THe next level would be to still assume that you stop at the same > >> place (i.e. you do not support patches from multiple series in the > >> same mailbox), but use the number-santized-subject format. This > >> would be a bit more involved, but I think all you need to update on > >> the "am" side is where the am_run() assigns the message file to the > >> local variable "mail". You know the temporary directory where you > >> told "mailsplit" to create these individual messages, so you should > >> be able to "opendir/readdir/closedir" and create a list of numbered > >> files in the directory very early in "git am". Knowing msgnum(state) > >> at that point in the loop, it should be trivial to change the code > >> that currently assumes the 4-th file is named "0004" to check for > >> the file whose name begins with "0004-". > > > > Yea, we pretty much just have to get the git-am process to work with the > > new names. I can look at using opendir/readdir here instead. > > Not "here", but probably just after you called "mailsplit" and saw > it return. After that nobody should be adding more split mail > messages to the directory, so you do it once to grab all filenames. > > > Ya I don't care much about multiple series. I care more about making it > > handle scrambled series better than it does now. I download series off > > of lore.kernel.org (public-inbox based) and those seem to routinely have > > series out-of-order. I suspect this is because it bases them on arrival > > date and sometimes certain mailers get it out of order when sending. > > Yeah, and that is why I said it would be a better use of the > engineering resource to fix it at the source. Such a fix will > benefit folks with existing versions of "git am", not needing to > wait for your improved version. > > Thanks. I went and talked to the public-inbox folks, and discovered that there is a known problem and solution, with a utility called b4 intended for downloading mbox files from the public-inbox https://b4.docs.kernel.org/ Thought I'd mention that here if anyone else reading this thread was curious about an ultimate solution. b4 will find patches in the series, sort them, remove the replies and can do some other common cleanup operations including things like applying tags from other messages on the list.
diff --git a/Documentation/git-mailsplit.txt b/Documentation/git-mailsplit.txt index 3f0a6662c81e..2e5ba45e1988 100644 --- a/Documentation/git-mailsplit.txt +++ b/Documentation/git-mailsplit.txt @@ -9,6 +9,7 @@ SYNOPSIS -------- [verse] 'git mailsplit' [-b] [-f<nn>] [-d<prec>] [--keep-cr] [--mboxrd] + [--name-by-subject] -o<directory> [--] [(<mbox>|<Maildir>)...] DESCRIPTION @@ -52,6 +53,10 @@ OPTIONS Input is of the "mboxrd" format and "^>+From " line escaping is reversed. +--name-by-subject:: + Include the sanitized subject in the generated filenames, in + addition to the sequence number. + GIT --- Part of the linkgit:git[1] suite diff --git a/builtin/mailsplit.c b/builtin/mailsplit.c index 3af9ddb8ae5c..df81782d05b3 100644 --- a/builtin/mailsplit.c +++ b/builtin/mailsplit.c @@ -8,9 +8,10 @@ #include "gettext.h" #include "string-list.h" #include "strbuf.h" +#include "pretty.h" static const char git_mailsplit_usage[] = -"git mailsplit [-d<prec>] [-f<n>] [-b] [--keep-cr] -o<directory> [(<mbox>|<Maildir>)...]"; +"git mailsplit [-d<prec>] [-f<n>] [-b] [--keep-cr] [--name-by-subject] -o<directory> [(<mbox>|<Maildir>)...]"; static int is_from_line(const char *line, int len) { @@ -46,6 +47,7 @@ static int is_from_line(const char *line, int len) static struct strbuf buf = STRBUF_INIT; static int keep_cr; static int mboxrd; +static int name_by_subject; static int is_gtfrom(const struct strbuf *buf) { @@ -66,6 +68,9 @@ static int is_gtfrom(const struct strbuf *buf) */ static int split_one(FILE *mbox, const char *name, int allow_bare) { + struct strbuf sanitized_filename = STRBUF_INIT; + const char *subject_start; + size_t subject_len; FILE *output; int fd; int status = 0; @@ -101,10 +106,26 @@ static int split_one(FILE *mbox, const char *name, int allow_bare) } die_errno("cannot read mbox"); } + + /* Get a sanitized filename from the subject */ + if (name_by_subject && !sanitized_filename.len && + skip_prefix_mem(buf.buf, buf.len, "Subject:", + &subject_start, &subject_len)) { + strbuf_addf(&sanitized_filename, "%s-", name); + format_sanitized_subject(&sanitized_filename, + subject_start, + subject_len); + } + if (!is_bare && is_from_line(buf.buf, buf.len)) break; /* done with one message */ } fclose(output); + + if (name_by_subject && sanitized_filename.len) + rename(name, sanitized_filename.buf); + strbuf_release(&sanitized_filename); + return status; } @@ -296,6 +317,8 @@ int cmd_mailsplit(int argc, const char **argv, const char *prefix) usage(git_mailsplit_usage); } else if ( arg[1] == 'b' && !arg[2] ) { allow_bare = 1; + } else if (!strcmp(arg, "--name-by-subject")) { + name_by_subject = 1; } else if (!strcmp(arg, "--keep-cr")) { keep_cr = 1; } else if ( arg[1] == 'o' && arg[2] ) { diff --git a/t/t5100-mailinfo.sh b/t/t5100-mailinfo.sh index c8d06554541c..4826735c6033 100755 --- a/t/t5100-mailinfo.sh +++ b/t/t5100-mailinfo.sh @@ -44,6 +44,31 @@ do ' done +test_expect_success 'split sample box with --name-by-subject' ' + mkdir name-by-subject && + git mailsplit --name-by-subject -oname-by-subject "$DATA/sample.mbox" >last && + last=$(cat last) && + echo total is $last && + test $(cat last) = 18 +' + +check_mailinfo_name_by_subject () { + mail=$1 + mo="$(basename "$mail" | cut -c1-4)" + echo "$(basename "$mail")" >"sanitized$mo" && + git mailinfo -u "msg$mo" "patch$mo" <"$mail" >"info$mo" && + test_cmp "$DATA/msg$mo" "msg$mo" && + test_cmp "$DATA/patch$mo" "patch$mo" && + test_cmp "$DATA/info$mo" "info$mo" && + test_cmp "$DATA/sanitized$mo" "sanitized$mo" +} + +for mail in name-by-subject/00* +do + test_expect_success "check --name-by-subject $mail" ' + check_mailinfo_name_by_subject "$mail" + ' +done test_expect_success 'split box with rfc2047 samples' \ 'mkdir rfc2047 &&