# [PATCH v3] commit.c: Replace starts_with() with skip_prefix() 7 messages from 2014-03-06 to 2014-03-09. Participants: Karthik Nayak, Junio C Hamano, Eric Sunshine, karthik nayak. Thread: https://gitlist.dev/t/36075 ## Karthik Nayak, 2014-03-06 17:05 Subject: [PATCH v3] commit.c: Replace starts_with() with skip_prefix() Message-ID: <1394125521-9341-1-git-send-email-karthik.188@gmail.com> URL: https://gitlist.dev/e/1394125521-9341-1-git-send-email-karthik.188%40gmail.com ``` Replace all instances of starts_with() by skip_prefix(), which can not only be used to check presence of a prefix, but also used further on as it returns the string after the prefix, if the prefix is present. And also manages to do, what the current code does in two steps. Signed-off-by: Karthik Nayak --- Hello Eric, In this patch, I have: 1. Fixed the improper placement of buf_date , initialised to a NULL string. 2. Fixed whitespace. 3. Better naming as per your suggestion. 4. Fixed the initilisation before the if statement. Thanks for your suggestion. --- commit.c | 22 +++++++++++----------- 1 file changed, 11 insertions(+), 11 deletions(-) diff --git a/commit.c b/commit.c index 6bf4fe0..4144e00 100644 --- a/commit.c +++ b/commit.c @@ -553,6 +553,7 @@ static void record_author_date(struct author_date_slab *author_date, struct ident_split ident; char *date_end; unsigned long date; + const char *buf_date; if (!commit->buffer) { unsigned long size; @@ -565,15 +566,15 @@ static void record_author_date(struct author_date_slab *author_date, for (buf = commit->buffer ? commit->buffer : buffer; buf; buf = line_end + 1) { + buf_date = skip_prefix(buf, "author "); line_end = strchrnul(buf, '\n'); - if (!starts_with(buf, "author ")) { + if (!buf_date) { if (!line_end[0] || line_end[1] == '\n') return; /* end of header */ continue; } - if (split_ident_line(&ident, - buf + strlen("author "), - line_end - (buf + strlen("author "))) || + if (split_ident_line(&ident, buf_date, + line_end - buf_date) || !ident.date_begin || !ident.date_end) goto fail_exit; /* malformed "author" line */ break; @@ -1098,6 +1099,7 @@ int parse_signed_commit(const unsigned char *sha1, char *buffer = read_sha1_file(sha1, &type, &size); int in_signature, saw_signature = -1; char *line, *tail; + const char *gpg_sig; if (!buffer || type != OBJ_COMMIT) goto cleanup; @@ -1113,9 +1115,9 @@ int parse_signed_commit(const unsigned char *sha1, next = next ? next + 1 : tail; if (in_signature && line[0] == ' ') sig = line + 1; - else if (starts_with(line, gpg_sig_header) && - line[gpg_sig_header_len] == ' ') - sig = line + gpg_sig_header_len + 1; + else if ((gpg_sig = skip_prefix(line, gpg_sig_header)) + && gpg_sig[0] == ' ') + sig = gpg_sig + 1; if (sig) { strbuf_add(signature, sig, next - sig); saw_signature = 1; @@ -1193,10 +1195,8 @@ static void parse_gpg_output(struct signature_check *sigc) for (i = 0; i < ARRAY_SIZE(sigcheck_gpg_status); i++) { const char *found, *next; - if (starts_with(buf, sigcheck_gpg_status[i].check + 1)) { - /* At the very beginning of the buffer */ - found = buf + strlen(sigcheck_gpg_status[i].check + 1); - } else { + found = skip_prefix(buf, sigcheck_gpg_status[i].check + 1); + if (!found) { found = strstr(buf, sigcheck_gpg_status[i].check); if (!found) continue; -- 1.9.0.138.g2de3478 ``` ## Junio C Hamano, 2014-03-06 19:36 Subject: Re: [PATCH v3] commit.c: Replace starts_with() with skip_prefix() Message-ID: URL: https://gitlist.dev/e/xmqqlhwnxgo7.fsf%40gitster.dls.corp.google.com In-Reply-To: <1394125521-9341-1-git-send-email-karthik.188@gmail.com> ``` We already have 147972b1 (commit.c: use skip_prefix() instead of starts_with(), 2014-03-04) that covers the record_author_date() and parse_gpg_output(), don't we? ``` ## Junio C Hamano, 2014-03-06 19:42 Subject: Re: [PATCH v3] commit.c: Replace starts_with() with skip_prefix() Message-ID: URL: https://gitlist.dev/e/xmqqfvmvxgf5.fsf%40gitster.dls.corp.google.com In-Reply-To: <1394125521-9341-1-git-send-email-karthik.188@gmail.com> ``` Karthik Nayak writes: > @@ -1098,6 +1099,7 @@ int parse_signed_commit(const unsigned char *sha1, > char *buffer = read_sha1_file(sha1, &type, &size); > int in_signature, saw_signature = -1; > char *line, *tail; > + const char *gpg_sig; > > if (!buffer || type != OBJ_COMMIT) > goto cleanup; > @@ -1113,9 +1115,9 @@ int parse_signed_commit(const unsigned char *sha1, > next = next ? next + 1 : tail; > if (in_signature && line[0] == ' ') > sig = line + 1; > - else if (starts_with(line, gpg_sig_header) && > - line[gpg_sig_header_len] == ' ') > - sig = line + gpg_sig_header_len + 1; > + else if ((gpg_sig = skip_prefix(line, gpg_sig_header)) > + && gpg_sig[0] == ' ') > + sig = gpg_sig + 1; I am not sure if this hunk is a great improvement, as we know the length of what we are skipping in the gpg_sig_header_len constant that is used throughout this file. ``` ## Eric Sunshine, 2014-03-07 09:34 Subject: Re: [PATCH v3] commit.c: Replace starts_with() with skip_prefix() Message-ID: URL: https://gitlist.dev/e/CAPig%2BcR0tw4em_mPNLdp1-gNER2G1FyAVCyZAb9_TbtSyuGyaA%40mail.gmail.com In-Reply-To: <1394125521-9341-1-git-send-email-karthik.188@gmail.com> ``` On Thu, Mar 6, 2014 at 12:05 PM, Karthik Nayak wrote: > Replace all instances of starts_with() by skip_prefix(), > which can not only be used to check presence of a prefix, > but also used further on as it returns the string after the prefix, > if the prefix is present. And also manages to do, what the current > code does in two steps. Better. Thanks. > Signed-off-by: Karthik Nayak > --- > Hello Eric, > In this patch, I have: > 1. Fixed the improper placement of buf_date , initialised to a NULL string. > 2. Fixed whitespace. > 3. Better naming as per your suggestion. > 4. Fixed the initilisation before the if statement. > Thanks for your suggestion. > --- > commit.c | 22 +++++++++++----------- > 1 file changed, 11 insertions(+), 11 deletions(-) > > diff --git a/commit.c b/commit.c > index 6bf4fe0..4144e00 100644 > --- a/commit.c > +++ b/commit.c > @@ -553,6 +553,7 @@ static void record_author_date(struct author_date_slab *author_date, > struct ident_split ident; > char *date_end; > unsigned long date; > + const char *buf_date; > > if (!commit->buffer) { > unsigned long size; > @@ -565,15 +566,15 @@ static void record_author_date(struct author_date_slab *author_date, > for (buf = commit->buffer ? commit->buffer : buffer; > buf; > buf = line_end + 1) { > + buf_date = skip_prefix(buf, "author "); The data after "author " is identification information (name, email), not date. In fact, this information gets passed to the function split_ident_line(), so a better name for this variable is 'ident_line' (but not the misspelling 'indent_line' from one of your earlier attempts). > line_end = strchrnul(buf, '\n'); > - if (!starts_with(buf, "author ")) { > + if (!buf_date) { > if (!line_end[0] || line_end[1] == '\n') > return; /* end of header */ > continue; > } > - if (split_ident_line(&ident, > - buf + strlen("author "), > - line_end - (buf + strlen("author "))) || > + if (split_ident_line(&ident, buf_date, > + line_end - buf_date) || > !ident.date_begin || !ident.date_end) > goto fail_exit; /* malformed "author" line */ > break; > @@ -1098,6 +1099,7 @@ int parse_signed_commit(const unsigned char *sha1, > char *buffer = read_sha1_file(sha1, &type, &size); > int in_signature, saw_signature = -1; > char *line, *tail; > + const char *gpg_sig; > > if (!buffer || type != OBJ_COMMIT) > goto cleanup; > @@ -1113,9 +1115,9 @@ int parse_signed_commit(const unsigned char *sha1, > next = next ? next + 1 : tail; > if (in_signature && line[0] == ' ') > sig = line + 1; > - else if (starts_with(line, gpg_sig_header) && > - line[gpg_sig_header_len] == ' ') > - sig = line + gpg_sig_header_len + 1; > + else if ((gpg_sig = skip_prefix(line, gpg_sig_header)) > + && gpg_sig[0] == ' ') > + sig = gpg_sig + 1; Other than the poor variable name 'buf_date' and the review comments by Junio and Tanay that this particular change is of questionable value and perhaps should be dropped, this version of the patch looks reasonable. Thanks. > if (sig) { > strbuf_add(signature, sig, next - sig); > saw_signature = 1; > @@ -1193,10 +1195,8 @@ static void parse_gpg_output(struct signature_check *sigc) > for (i = 0; i < ARRAY_SIZE(sigcheck_gpg_status); i++) { > const char *found, *next; > > - if (starts_with(buf, sigcheck_gpg_status[i].check + 1)) { > - /* At the very beginning of the buffer */ > - found = buf + strlen(sigcheck_gpg_status[i].check + 1); > - } else { > + found = skip_prefix(buf, sigcheck_gpg_status[i].check + 1); > + if (!found) { > found = strstr(buf, sigcheck_gpg_status[i].check); > if (!found) > continue; > -- > 1.9.0.138.g2de3478 > ``` ## karthik nayak, 2014-03-07 12:39 Subject: Re: [PATCH v3] commit.c: Replace starts_with() with skip_prefix() Message-ID: URL: https://gitlist.dev/e/CAOLa%3DZQZOxEzXUEwCVpesNzRhJ5_sMbtbUpaYjdfEOOrakbG0Q%40mail.gmail.com In-Reply-To: ``` Hello Eric, Thanks for your reply, and for that information. should i patch again or this should do? And what next? Talk to the mentor? Thanks ``` ## Eric Sunshine, 2014-03-09 07:49 Subject: Re: [PATCH v3] commit.c: Replace starts_with() with skip_prefix() Message-ID: URL: https://gitlist.dev/e/CAPig%2BcS9Eh7bde90SsVWqeWP%2BfZqDTV3fk6FryoP5SFg-FyfTw%40mail.gmail.com In-Reply-To: ``` On Fri, Mar 7, 2014 at 5:49 AM, karthik nayak wrote: > Hello Eric, > Thanks for your reply, and for that information. should i patch again or > this should do? > And what next? Talk to the mentor? The ultimate authority deciding if a patch is ready is Junio, as it would have to be accepted into his tree. Since he already accepted a similar patch from a different potential GSoC applicant, it may not make sense to refine this one further. What is important is that you are now familiar with the review process on this project, and the mentors (hopefully) have gained insight into your abilities and how you interact with reviewers (which was the goal of these microprojects). Probably best at this point is to consider a proposed project [1] or choose your own, and start the task of applying for a GSoC position (by whatever means that is done). [1]: https://github.com/git/git.github.io/blob/master/SoC-2014-Ideas.md > On Fri, Mar 7, 2014 at 3:04 PM, Eric Sunshine > wrote: >> >> On Thu, Mar 6, 2014 at 12:05 PM, Karthik Nayak >> wrote: >> > Replace all instances of starts_with() by skip_prefix(), >> > which can not only be used to check presence of a prefix, >> > but also used further on as it returns the string after the prefix, >> > if the prefix is present. And also manages to do, what the current >> > code does in two steps. >> >> Better. Thanks. >> >> > Signed-off-by: Karthik Nayak >> > --- >> > Hello Eric, >> > In this patch, I have: >> > 1. Fixed the improper placement of buf_date , initialised to a NULL >> > string. >> > 2. Fixed whitespace. >> > 3. Better naming as per your suggestion. >> > 4. Fixed the initilisation before the if statement. >> > Thanks for your suggestion. >> > --- >> > commit.c | 22 +++++++++++----------- >> > 1 file changed, 11 insertions(+), 11 deletions(-) >> > >> > diff --git a/commit.c b/commit.c >> > index 6bf4fe0..4144e00 100644 >> > --- a/commit.c >> > +++ b/commit.c >> > @@ -553,6 +553,7 @@ static void record_author_date(struct >> > author_date_slab *author_date, >> > struct ident_split ident; >> > char *date_end; >> > unsigned long date; >> > + const char *buf_date; >> > >> > if (!commit->buffer) { >> > unsigned long size; >> > @@ -565,15 +566,15 @@ static void record_author_date(struct >> > author_date_slab *author_date, >> > for (buf = commit->buffer ? commit->buffer : buffer; >> > buf; >> > buf = line_end + 1) { >> > + buf_date = skip_prefix(buf, "author "); >> >> The data after "author " is identification information (name, email), >> not date. In fact, this information gets passed to the function >> split_ident_line(), so a better name for this variable is 'ident_line' >> (but not the misspelling 'indent_line' from one of your earlier >> attempts). >> >> > line_end = strchrnul(buf, '\n'); >> > - if (!starts_with(buf, "author ")) { >> > + if (!buf_date) { >> > if (!line_end[0] || line_end[1] == '\n') >> > return; /* end of header */ >> > continue; >> > } >> > - if (split_ident_line(&ident, >> > - buf + strlen("author "), >> > - line_end - (buf + strlen("author >> > "))) || >> > + if (split_ident_line(&ident, buf_date, >> > + line_end - buf_date) || >> > !ident.date_begin || !ident.date_end) >> > goto fail_exit; /* malformed "author" line */ >> > break; >> > @@ -1098,6 +1099,7 @@ int parse_signed_commit(const unsigned char *sha1, >> > char *buffer = read_sha1_file(sha1, &type, &size); >> > int in_signature, saw_signature = -1; >> > char *line, *tail; >> > + const char *gpg_sig; >> > >> > if (!buffer || type != OBJ_COMMIT) >> > goto cleanup; >> > @@ -1113,9 +1115,9 @@ int parse_signed_commit(const unsigned char *sha1, >> > next = next ? next + 1 : tail; >> > if (in_signature && line[0] == ' ') >> > sig = line + 1; >> > - else if (starts_with(line, gpg_sig_header) && >> > - line[gpg_sig_header_len] == ' ') >> > - sig = line + gpg_sig_header_len + 1; >> > + else if ((gpg_sig = skip_prefix(line, gpg_sig_header)) >> > + && gpg_sig[0] == ' ') >> > + sig = gpg_sig + 1; >> >> Other than the poor variable name 'buf_date' and the review comments >> by Junio and Tanay that this particular change is of questionable >> value and perhaps should be dropped, this version of the patch looks >> reasonable. >> >> Thanks. >> >> > if (sig) { >> > strbuf_add(signature, sig, next - sig); >> > saw_signature = 1; >> > @@ -1193,10 +1195,8 @@ static void parse_gpg_output(struct >> > signature_check *sigc) >> > for (i = 0; i < ARRAY_SIZE(sigcheck_gpg_status); i++) { >> > const char *found, *next; >> > >> > - if (starts_with(buf, sigcheck_gpg_status[i].check + 1)) >> > { >> > - /* At the very beginning of the buffer */ >> > - found = buf + >> > strlen(sigcheck_gpg_status[i].check + 1); >> > - } else { >> > + found = skip_prefix(buf, sigcheck_gpg_status[i].check + >> > 1); >> > + if (!found) { >> > found = strstr(buf, >> > sigcheck_gpg_status[i].check); >> > if (!found) >> > continue; >> > -- >> > 1.9.0.138.g2de3478 >> > > > ``` ## karthik nayak, 2014-03-09 11:10 Subject: Re: [PATCH v3] commit.c: Replace starts_with() with skip_prefix() Message-ID: URL: https://gitlist.dev/e/CAOLa%3DZQWREUTSnbm5-XM2yZLR3HXBEkENrHaOPzDFwDx2S3L%3Dg%40mail.gmail.com In-Reply-To: ``` Hey Eric, Its been nice learning from you about how to submit patches to git. was a nice learning curve, now I'm looking into the ideas and will contact the appropriate mentor soon with a plan. Thanks - Karthik On Sun, Mar 9, 2014 at 1:19 PM, Eric Sunshine wrote: > On Fri, Mar 7, 2014 at 5:49 AM, karthik nayak wrote: >> Hello Eric, >> Thanks for your reply, and for that information. should i patch again or >> this should do? >> And what next? Talk to the mentor? > > The ultimate authority deciding if a patch is ready is Junio, as it > would have to be accepted into his tree. Since he already accepted a > similar patch from a different potential GSoC applicant, it may not > make sense to refine this one further. What is important is that you > are now familiar with the review process on this project, and the > mentors (hopefully) have gained insight into your abilities and how > you interact with reviewers (which was the goal of these > microprojects). > > Probably best at this point is to consider a proposed project [1] or > choose your own, and start the task of applying for a GSoC position > (by whatever means that is done). > > [1]: https://github.com/git/git.github.io/blob/master/SoC-2014-Ideas.md > >> On Fri, Mar 7, 2014 at 3:04 PM, Eric Sunshine >> wrote: >>> >>> On Thu, Mar 6, 2014 at 12:05 PM, Karthik Nayak >>> wrote: >>> > Replace all instances of starts_with() by skip_prefix(), >>> > which can not only be used to check presence of a prefix, >>> > but also used further on as it returns the string after the prefix, >>> > if the prefix is present. And also manages to do, what the current >>> > code does in two steps. >>> >>> Better. Thanks. >>> >>> > Signed-off-by: Karthik Nayak >>> > --- >>> > Hello Eric, >>> > In this patch, I have: >>> > 1. Fixed the improper placement of buf_date , initialised to a NULL >>> > string. >>> > 2. Fixed whitespace. >>> > 3. Better naming as per your suggestion. >>> > 4. Fixed the initilisation before the if statement. >>> > Thanks for your suggestion. >>> > --- >>> > commit.c | 22 +++++++++++----------- >>> > 1 file changed, 11 insertions(+), 11 deletions(-) >>> > >>> > diff --git a/commit.c b/commit.c >>> > index 6bf4fe0..4144e00 100644 >>> > --- a/commit.c >>> > +++ b/commit.c >>> > @@ -553,6 +553,7 @@ static void record_author_date(struct >>> > author_date_slab *author_date, >>> > struct ident_split ident; >>> > char *date_end; >>> > unsigned long date; >>> > + const char *buf_date; >>> > >>> > if (!commit->buffer) { >>> > unsigned long size; >>> > @@ -565,15 +566,15 @@ static void record_author_date(struct >>> > author_date_slab *author_date, >>> > for (buf = commit->buffer ? commit->buffer : buffer; >>> > buf; >>> > buf = line_end + 1) { >>> > + buf_date = skip_prefix(buf, "author "); >>> >>> The data after "author " is identification information (name, email), >>> not date. In fact, this information gets passed to the function >>> split_ident_line(), so a better name for this variable is 'ident_line' >>> (but not the misspelling 'indent_line' from one of your earlier >>> attempts). >>> >>> > line_end = strchrnul(buf, '\n'); >>> > - if (!starts_with(buf, "author ")) { >>> > + if (!buf_date) { >>> > if (!line_end[0] || line_end[1] == '\n') >>> > return; /* end of header */ >>> > continue; >>> > } >>> > - if (split_ident_line(&ident, >>> > - buf + strlen("author "), >>> > - line_end - (buf + strlen("author >>> > "))) || >>> > + if (split_ident_line(&ident, buf_date, >>> > + line_end - buf_date) || >>> > !ident.date_begin || !ident.date_end) >>> > goto fail_exit; /* malformed "author" line */ >>> > break; >>> > @@ -1098,6 +1099,7 @@ int parse_signed_commit(const unsigned char *sha1, >>> > char *buffer = read_sha1_file(sha1, &type, &size); >>> > int in_signature, saw_signature = -1; >>> > char *line, *tail; >>> > + const char *gpg_sig; >>> > >>> > if (!buffer || type != OBJ_COMMIT) >>> > goto cleanup; >>> > @@ -1113,9 +1115,9 @@ int parse_signed_commit(const unsigned char *sha1, >>> > next = next ? next + 1 : tail; >>> > if (in_signature && line[0] == ' ') >>> > sig = line + 1; >>> > - else if (starts_with(line, gpg_sig_header) && >>> > - line[gpg_sig_header_len] == ' ') >>> > - sig = line + gpg_sig_header_len + 1; >>> > + else if ((gpg_sig = skip_prefix(line, gpg_sig_header)) >>> > + && gpg_sig[0] == ' ') >>> > + sig = gpg_sig + 1; >>> >>> Other than the poor variable name 'buf_date' and the review comments >>> by Junio and Tanay that this particular change is of questionable >>> value and perhaps should be dropped, this version of the patch looks >>> reasonable. >>> >>> Thanks. >>> >>> > if (sig) { >>> > strbuf_add(signature, sig, next - sig); >>> > saw_signature = 1; >>> > @@ -1193,10 +1195,8 @@ static void parse_gpg_output(struct >>> > signature_check *sigc) >>> > for (i = 0; i < ARRAY_SIZE(sigcheck_gpg_status); i++) { >>> > const char *found, *next; >>> > >>> > - if (starts_with(buf, sigcheck_gpg_status[i].check + 1)) >>> > { >>> > - /* At the very beginning of the buffer */ >>> > - found = buf + >>> > strlen(sigcheck_gpg_status[i].check + 1); >>> > - } else { >>> > + found = skip_prefix(buf, sigcheck_gpg_status[i].check + >>> > 1); >>> > + if (!found) { >>> > found = strstr(buf, >>> > sigcheck_gpg_status[i].check); >>> > if (!found) >>> > continue; >>> > -- >>> > 1.9.0.138.g2de3478 >>> > >> >> ```