[PATCH 2 of 3 V4 RFC] model (1): c-hglib: hg_log() level 1 function

Giovanni Gherdovich g.gherdovich at gmail.com
Wed Sep 4 04:13:40 CDT 2013


Hi Iulian,

this is a long review.

I add a table of contents so to make random access easier.

1. "#define _GNU_SOURCE", wat
2. log template is better changed
3. Rename the 'adding_data()' function
4. Rename "new_cset_size" and "cset" local variable in 'erase_cset()'
5. Check correct usage of free(3)
6. No reason why hg_log() should return a hg_csetstream_buffer*
7. Discuss return value of hg_fetch_cset_entry()
8. Do we really need to expose hg_csetstream_buffer* to the user code
9. Review your logic in hg_fetch_cset_entry(), make it more "linear"
10. hg_log() should handle error message, too
11. strlen(3) is not a null byte detector, don't use it as such
12. hg_exitcode() to be called by user code, not inside
hg_fetch_cset_entry()


:::: diff --git a/client.c b/client.c
:::: new file mode 100644
:::: --- /dev/null
:::: +++ b/client.c
:::: @@ -0,0 +1,175 @@
:::: +#define _GNU_SOURCE

I would prefer not to see the _GNU_SOURCE macro identifier defined in your
source.
Either make your code work without it, or provide a thorough explanation of
why
it is so desperately needed.

:::: +#include <errno.h>
:::: +#include <fcntl.h>
:::: +#include <stdlib.h>
:::: +#include <stdio.h>
:::: +#include <string.h>
:::: +#include <unistd.h>
:::: +#include <sys/types.h>
:::: +#include <sys/wait.h>
:::: +#include <signal.h>
:::: +
:::: +
:::: +#include "client.h"
:::: +#include "utils.h"
:::: +
:::: +#define HGPATH "hg"
:::: +#define CHANGESET "\\0{rev}\\n{node}\\n{tags}\\n{branch}\\n{author}\
:::: +                        \\n{date|isodate}\\n{desc}"

As I write below, I think that hg_fetch_cset_entry() is a little less
cumbersome to implement if the '\0' is at the end of the changeset data
instead of being at the beginning.

Reason:
in either case, one changeset of your history has to be "special cased".
If the '\0' marker is at the beginning, the last changeset has to be treated
exceptionally (where will it end?).
If the '\0' marker is at the end, the first changeset has to be treated
exceptionally (where will it begin?).
But in the latter case, you know the answer, since the changeset data begins
right after you issue the command 'hg log' to cmdsrv.

We chosed the above template since it's the same as in JavaHG at commit
https://bitbucket.org/aragost/javahg/commits/7855d21c99e1#chg-src/main/java/com/aragost/javahg/Changeset.java
But I went checking again, and it looks like they ended up with '\0' at the
end:
https://bitbucket.org/aragost/javahg/src/tip/src/main/resources/styles/changesets.style?at=default

Martin Geisler please correct me if I'm wrong.

:::: +
:::: +
:::: +
:::: +/**
:::: + * \brief 'Parse a changeset'. It's more like pointing to the correct
position.
:::: + *
:::: + * The changeset could be found on buff pointer. To not duplicate the
data I
:::: + * choose to point every hg_cset_entry field to the right position.
:::: + * \param cset The pointer where changeset could be found.
:::: + * \param ce   The hg_cset_entry structure where the changeset will
be parse.
:::: + * \retval 0 if successful.
:::: + * */
:::: +int parse_changeset(char *cset, hg_cset_entry *ce)
:::: +{
:::: +    char *position = cset;
:::: +    /* set pointer for revision position */
:::: +    ce->rev = cset;
:::: +    position = strstr(position, "\n");
:::: +    cset[position - cset] = '\0';
:::: +
:::: +    /* set pointer for node position */
:::: +    ce->node = position + 1;
:::: +    position = strstr(position + 1, "\n");
:::: +    cset[position - cset] = '\0';
:::: +
:::: +    /* set pointer for tag position */
:::: +    ce->tags = position + 1;
:::: +    position = strstr(position + 1, "\n");
:::: +    cset[position - cset] = '\0';
:::: +
:::: +    /* set pointer for branch position */
:::: +    ce->branch = position + 1;
:::: +    position = strstr(position + 1, "\n");
:::: +    cset[position - cset] = '\0';
:::: +
:::: +    /* set pointer for author position */
:::: +    ce->author = position + 1;
:::: +    position = strstr(position + 1, "\n");
:::: +    cset[position - cset] = '\0';
:::: +
:::: +    /* set pointer for data position */
:::: +    ce->date = position + 1;
:::: +    position = strstr(position + 1, "\n");
:::: +    cset[position - cset] = '\0';
:::: +
:::: +    /* set pointer for description position */
:::: +    ce->desc = position + 1;
:::: +    /* */
:::: +    return 0;
:::: +}
:::: +
:::: +/* Adding to the destination pointer the source pointer. */
:::: +int adding_data(char **dest, char *source, int dsize, int ssize)

Rename to 'append_data'. Please don't use the gerund tense in function
names,
the simple present tense is better.

:::: +{
:::: +    if(*dest == NULL){
:::: +        *dest = malloc(ssize + 1);
:::: +        memcpy(*dest, source, ssize + 1);
:::: +    } else {
:::: +        *dest = realloc(*dest, dsize + ssize + 2);
:::: +        memcpy(*dest + dsize, source, ssize + 1);
:::: +    }
:::: +    return 0;
:::: +}
:::: +
:::: +/* Erase the top cset from cset pointer. */
:::: +int erase_cset(char **cset, int buf_size, int first_cset_size)
:::: +{
:::: +    int new_cset_size = buf_size - first_cset_size;

(*) "new_cset_size" it is *not* the size of a changeset, it's the size of
the buffer.
    Name it accordingly, like "new_buf_size".
(*) 'cset' is not changeset data, is the content of a raw unparsed buffer.
    Name it accordingly.

:::: +    char *new_cset = malloc(new_cset_size + 1);
:::: +    memcpy(new_cset, *cset + first_cset_size, new_cset_size + 1);
:::: +    free(*cset);

The *cset you're free'ing here was malloc'ed in the 'adding_data' function,
as far as I can tell.
There you where appending a chunk of the input stream to your internal
buffer;
this chunk could possibly contain more than one changeset segment.

Please convince me that here you're not free'ing data below the "first
changeset" boundary,
because it looks very much like you are doing so.

:::: +    *cset = new_cset;
:::: +    return new_cset_size;
:::: +}
:::: +
:::: +
:::: +/* The high level log command for hglib API. */
:::: +hg_csetstream_buffer *hg_log(hg_handle *handle, char *option[])
:::: +{

I really do not understand why hg_log() has to return a
hg_csetstream_buffer*.
There is no special information necessary to initialize the
hg_csetstream_buffer*
that is available only from inside hg_log().
It is basically just the hg_handle, which is all over the place.

And, as I write a few lines below, I think hg_csetstream_buffer* is an
internal
that should not be exposed to the user code.

:::: +    hg_csetstream_buffer *cbuf = malloc(sizeof(hg_csetstream_buffer));
:::: +    cbuf->handle = handle;
:::: +
:::: +    cbuf->command = cmdbuilder("log", option, "--template", CHANGESET,
:::: +                            NULL);
:::: +
:::: +    if(hg_rawcommand(handle, cbuf->command) < 0){
:::: +        return NULL;
:::: +    }
:::: +
:::: +    cbuf->buffer = NULL;
:::: +    cbuf->buf_size = 0;
:::: +    cbuf->is_send = 0;
:::: +
:::: +    return cbuf;
:::: +}
:::: +
:::: +/* The cbuf next step. Getting the next changeset. */
:::: +int hg_fetch_cset_entry(hg_csetstream_buffer *cbuf, hg_cset_entry
*centry)
:::: +{

Let's look at things from 100 kilometers away:
here our goal is to write a function that enables the pattern

while( hg_fetch_cset_entry( cset_entry ) )
        { /* ...do stuff with cset_entry... */ }

cset_entry is "something" that contains info about *one*
and *only one* changeset.
The return value of hg_fetch_cset_entry() should be such that

(*) it's bool eval-ed to 1 if there is stuff to fetch,
(*) it's bool eval-ed to 0 if there is nothing more to fetch,
(*) and it can give clues if something goes wrong.

Pretty much like read(2).

Looking at your signature, I see two "output parameters":
(1) hg_csetstream_buffer *cbuf
(2) hg_cset_entry *centry

Now, the pointer cbuf look *a lot* like something
the user could not care less about. It's an "internal".

I wander if we can get rid of it, i.e. do not expose it on the function
prototype.
I *do understand* that the only reason we keep cbuf out of
hg_fetch_cset_entry()
is because we need it to be "persistent" across multiple function calls...
I just don't like it and wander if it can be done better and cleaner.

If, on the other side, I am wrong and cbuf is something the user code
will need at some time (but I doubt it), then hg_fetch_cset_entry()
is producing two different outputs (cbuf and centry), which is a clue
that probably it's doing too much.

:::: +    hg_header head = hg_head(cbuf->handle);
:::: +    int exitcode;
:::: +    char *get_data;
:::: +    int read_size;
:::: +
:::: +    /* Erase the first cset from cset pointer.
:::: +     * This cset was already pass to user.*/
:::: +    if(cbuf->is_send && cbuf->buf_size){
:::: +        cbuf->buf_size = erase_cset(&cbuf->buffer, cbuf->buf_size,
:::: +                        cbuf->first_cset_size);
:::: +    }

You start by cleaning up. I don't get it.
Can't you do this right after you parse the stream segment that correspond
to a full changeset?
I think the field hg_csetstream_buffer.is_send (which should be spelled
'is_sent', btw)
is useless and confusing.
If you clean up right after you use the data, no need to keep track of what
is sent or not.

More generally, I think the whole function needs a rewrite.
It's convoluted, difficult to follow, the intent is not clear.
To an external reader it looks like everytime you encountered a bug
you added an 'if-then' statement here and there.
At a first glance, I could not even say if all conditions are covered
(if you put an 'if-then' without an 'else', you're likely to be looking for
troubles).

I suggest you rewrite the body with something like:

if (buffer contains null byte){
        consume data;
        return something;
} else {
        read until you get a null byte;
}

(the above works if the null byte '\0' is at the end of the changeset
template.

:::: +    while(head.channel != 'r'){

I already told you this, just a reminder: you're forgetting the error
channel 'e' here.
Use the callback strategy to handle error messages.

:::: +        /* If there is a cset in cset pointer, then parse it and send
:::: +         * it to user.*/
:::: +        if(cbuf->buffer && strlen(cbuf->buffer + 1) < cbuf->buf_size
-1){

I *absolutely* don't like the use of strlen as a "null-byte" detector.
As you can imagine, strlen keeps a local pointer that runs until it finds
'\0':
https://sourceware.org/git/?p=glibc.git;a=blob;f=string/strlen.c
Since there is no guarantee that in your case the '\0' will ever be found,
this pointer can go a long way before strlen will return a value.

Please look for '\0' explicitely, like

-- -- >8 -- -- cut here -- -- >8 -- -- >8
char *char_ptr;
int null_byte_found = 0;
size_t cset_size = 0;

for (char_ptr = cbuf->buffer; char_ptr < cbuf->buffer + cbuf->buf_size;
++char_ptr)
        if (*char_ptr == '\0')
                int null_byte_found = 1;
                break;

if (null_byte_found)
        cset_size = char_ptr - cbuf->buffer;
-- -- >8 -- -- cut here -- -- >8 -- -- >8

:::: +            cbuf->first_cset_size = strlen(cbuf->buffer + 1) + 1;
:::: +            parse_changeset(cbuf->buffer + 1, centry);
:::: +            cbuf->is_send = 1;
:::: +            return head.length;

Here you return head.length, which makes very little sense:
it's just the lenght of the last 'o' chunk you got from cmdsrv,
while the parsed changeset you're returning (written in *centry)
might have been built up by multiple 'o' chunks one sticked after the other.

Another thing concerns me (we've discussed it over IRC, just reporting here
for posterity): in your current hg_rawread implementation, if cmdsrv sends
you a chunk bigger than the size you're willing to read (i.e. you read in
4K chunks),
what you do is *writing* the amount of bytes still to be read into
hg_header.length
(see
http://thread.gmane.org/gmane.comp.version-control.mercurial.devel/62348/focus=62351)

So: either we find a way to "const-ize" all variables of type hg_header,
either we don't refer to those values out of the blue when various
side-effects
are playing jokes under the hood.

:::: +        }
:::: +        else{
:::: +            /* Getting the next data from cmdserver and put on the
:::: +             * end of the cset pointer. */
:::: +            get_data = malloc(head.length + 1);
:::: +            if(read_size = hg_rawread(cbuf->handle, get_data,
:::: +                        head.length), read_size < 0){
:::: +                return -1;
:::: +            }
:::: +            adding_data(&cbuf->buffer, get_data, cbuf->buf_size,
:::: +                                read_size);
:::: +            cbuf->buf_size += read_size;
:::: +            head = hg_head(cbuf->handle);
:::: +            free(get_data);
:::: +        }
:::: +    }
:::: +    /* After, receiveing the last message, there still could be some
:::: +     * csets on cset pointer. */
:::: +    if(cbuf->buffer && strlen(cbuf->buffer + 1) == cbuf->buf_size -1){

A rather obscure condition to say
"in the first (cbuf->buf_size) bytes I don't have any \0 character".
Please make things more explicit, and please don't use strlen as "null byte
detector".


:::: +        cbuf->first_cset_size = strlen(cbuf->buffer + 1) + 1;
:::: +        parse_changeset(cbuf->buffer + 1, centry);
:::: +        cbuf->buf_size = 0;
:::: +        cbuf->is_send = 0;
:::: +        return head.length;
:::: +    /* Parse first cset from the remaining data. */
:::: +    }else if(cbuf->buf_size && cbuf->is_send){
:::: +        cbuf->first_cset_size = strlen(cbuf->buffer + 1) + 1;
:::: +        parse_changeset(cbuf->buffer + 1, centry);
:::: +        cbuf->is_send = 1;
:::: +        return head.length;
:::: +    }
:::: +
:::: +    exitcode = hg_exitcode(cbuf->handle);
:::: +    free(cbuf->command);
:::: +    free(cbuf->buffer);
:::: +    free(cbuf);
:::: +    return exitcode;

Ok this must be a leftover from previous iterations; a few lines above
you're
returning a byte length, here you return a status code. More discipline
please.
More over, hg_exitcode() will be called by user code, not by this function.

:::: +
:::: +}
:::: diff --git a/client.h b/client.h
:::: --- a/client.h
:::: +++ b/client.h
:::: @@ -69,6 +69,24 @@
::::      char *out_data;
::::  } hg_handle;
::::
:::: +typedef struct hg_csetstream_buffer{
:::: +    hg_handle *handle;
:::: +    char **command;
:::: +    char *buffer;
:::: +    int buf_size;
:::: +    int is_send;

After my remark above, I don't think you need this flag.
It should be spelled "is_sent", anyway (note the `t`).

:::: +    int first_cset_size;
:::: +}hg_csetstream_buffer;
:::: +
:::: +typedef struct hg_cset_entry{
:::: +    char *author;
:::: +    char *branch;
:::: +    char *date;
:::: +    char *desc;
:::: +    char *node;
:::: +    char *rev;
:::: +    char *tags;
:::: +}hg_cset_entry;
::::
::::  /**
::::   * \brief Open the connection with the mercurial command server.
:::: @@ -215,4 +233,63 @@
::::   * */
::::  int hg_exitcode(hg_handle *handle);
::::
:::: +/**
:::: + * \brief hg_log command for hglib API.
:::: + *
:::: + * It's an advance function to get revision history. It's more like
the start
:::: + * point of the action, this function will prepare the query question
and will
:::: + * send it to the cmd-server.
:::: + *
:::: + * Return the revision history of the specified files or the entire
project.
:::: + * File history is shown without following rename or copy history of
files.
:::: + * Use follow with a filename to follow history across renames and
copies.
:::: + * follow without a filename will only show ancestors or descendants
of the
:::: + * starting revision. followfirst only follows the first parent of
merge
:::: + * revisions.
:::: + *
:::: + * If revrange isn't specified, the default is "tip:0" unless follow
is set,
:::: + * in which case the working directory parent is used as the starting
:::: + * revision.
:::: + *
:::: + * \param handle The handle of the connection, wherewith I want to
communicate
:::: + * \param option The option list for mercurial log command.
:::: + * \retval hg_csetstream_buffer A pointer to hg_csetstream_buffer
structure if
:::: + *                              successful
:::: + * \retval NULL to indicate an error, with errno set appropriately.
:::: + *
:::: + * errno can be:
:::: + *      - hg_rawcommand errors
:::: + * */
:::: +hg_csetstream_buffer *hg_log(hg_handle *handle, char *option[]);
:::: +
:::: +/**
:::: + * \brief The iterator step. Getting the next changeset.
:::: + *
:::: + * The revision history could have a huge mass of data. You cannot
pass the
:::: + * entire  history in one call, so we use an iterator-like mechanism.
Calling
:::: + * the hg_fetch_log_entry. The next changeset will be read from
cmd-server,
:::: + * parse and pass to hg_cset_entry structure.
:::: + * The cset_entry structure will handle a  changeset with the
following string
:::: + * fields:
:::: + *         - rev
:::: + *         - node
:::: + *         - tags (space delimited)
:::: + *         - branch
:::: + *         - author
:::: + *         - desc
:::: + *
:::: + * \param hg_csetstream_buffer The buffer structure to store cset
data.
:::: + * \param centry The hg_cset_entry structure where the changeset will
be stored
:::: + *               and pass
:::: + * \retval number The lenght for the pass changeset.
:::: + * \retval exitcode To indicate the end of current_command.
:::: + * \retval   -1 to indicate an error, with errno set appropriately.
:::: + *
:::: + * errno can be:
:::: + *      - EINVAL  - Invalid argument (handle it's set to a null
pointer)
:::: + *      - read(2) command errors
:::: + *      - read_header error
:::: + * */
:::: +int hg_fetch_cset_entry(hg_csetstream_buffer *cbuf, hg_cset_entry
*centry);
:::: +
::::  #endif
:::: diff --git a/main.c b/main.c
:::: new file mode 100644
:::: --- /dev/null
:::: +++ b/main.c
:::: @@ -0,0 +1,128 @@
:::: +#include <stdio.h>
:::: +#include <stdlib.h>
:::: +#include <string.h>
:::: +
:::: +#include <sys/wait.h>
:::: +#include <sys/types.h>
:::: +#include <unistd.h>
:::: +#include <sys/stat.h>
:::: +#include <fcntl.h>
:::: +
:::: +#include "client.h"
:::: +#include "utils.h"
:::: +
:::: +#define INIT_REPO  "init_test_repo"
:::: +
:::: +/****** Convenience functions. *******/
:::: +
:::: +/**
:::: + * \brief Create and setup the tmp directory where the acction will
happends.
:::: + * */
:::: +void setup_tmp()
:::: +{
:::: +    system("hg init tmp");
:::: +    chdir("tmp");
:::: +}
:::: +
:::: +/**
:::: + * \brief Remove the tmp directory and all his files.
:::: + * */
:::: +void clean_tmp()
:::: +{
:::: +    chdir("..");
:::: +    system("rm -rf tmp");
:::: +}
:::: +
:::: +/**
:::: + * \brief Fill the current repository with commits for log command.
:::: + * */
:::: +void setup_log()
:::: +{
:::: +    system("touch foo ; hg add foo ; hg commit -m 'foo file'");
:::: +    system("echo baloo > foo ; hg commit -m 'baloo text'");
:::: +    system("touch voodoo ; hg add voodoo ; hg commit -m voodoo");
:::: +    system("echo voodoo > voodoo ; hg commit -m 'voodoo text'");
:::: +}
:::: +
:::: +/******* Examples using level 1 implementations. ******/
:::: +
:::: +/**
:::: + * \brief Log command example.
:::: + *
:::: + * \param handle The handle of the connection, wherewith I want to
communicate
:::: + * \retval exitcode
:::: + * */
:::: +int hg_log_example(hg_handle *handle)
:::: +{
:::: +    char *option[] = {"-v", NULL};
:::: +    int nc;
:::: +
:::: +    /* hg_log function will a iterator. */
:::: +    hg_csetstream_buffer *log_iterator = hg_log(handle, option);
:::: +
:::: +    /* you need to alloc some space for log_entry_t structure */
:::: +    hg_cset_entry *le = malloc(sizeof(hg_cset_entry));
:::: +
:::: +    /* Getting the next changeset using the iterator-like mechanism.
:::: +       Print the changest from log_entry structure.*/
:::: +    while(nc = hg_fetch_cset_entry(log_iterator, le), nc > 0){
:::: +        printf("rev = %s \n", le->rev);
:::: +        printf("node = %s \n", le->node);
:::: +        printf("tags = %s \n", le->tags);
:::: +        printf("branch = %s \n", le->branch);
:::: +        printf("author = %s \n", le->author);
:::: +        printf("date = %s \n", le->date);
:::: +        printf("desc = %s \n", le->desc);
:::: +        printf("\n");
:::: +    }
:::: +
:::: +    free(le);
:::: +    /* last call for hg_fetch_log_entry will pass the exitcode */
:::: +    return nc;
:::: +}
:::: +
:::: +/** \brief Printing the welcome message.
:::: + *
:::: + * Will print the options that you will have in this example.
:::: + **/
:::: +void print_select_case()
:::: +{
:::: +    printf("Select test case to run:\n");
:::: +    printf("1) log \n");
:::: +    printf("\n");
:::: +    printf("Your choice: ");
:::: +}
:::: +
:::: +
:::: +
:::: +/***** Main function. *******/
:::: +/**
:::: + * \brief The main function
:::: + * */
:::: +int main(int argc, char **argv)
:::: +{
:::: +    int select_case;
:::: +    hg_handle *handle;
:::: +
:::: +    print_select_case();
:::: +    scanf("%d", &select_case);
:::: +    if(select_case < 1 || select_case > 1){
:::: +        printf("Your choice is not an option...\n");
:::: +        return -1;
:::: +    }
:::: +
:::: +    switch(select_case){
:::: +        case 1:
:::: +            setup_tmp();
:::: +            setup_log();
:::: +            handle = hg_open(NULL, "");
:::: +
:::: +            hg_log_example(handle);
:::: +
:::: +            hg_close(&handle);
:::: +            clean_tmp();
:::: +            break;
:::: +    }
:::: +
:::: +    return 0;
:::: +}


cheers,
GGhh
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://selenic.com/pipermail/mercurial-devel/attachments/20130904/732a4d43/attachment.html>


More information about the Mercurial-devel mailing list