The top-level mailinfo() would instead punt when the code in the
deeper part of the callchain detects an unrecoverable error in the
input.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Instead of dying in convert_to_utf8(), just report an error and let
the callers handle it.  Between the two callers:

 - decode_header() silently punts when it cannot parse a broken
   RFC2047 encoded text (e.g. when it sees anything other than B or
   Q after it sees "=?<charset>") by jumping to release_return,
   returning the string it successfully parsed out so far, to the
   caller.  A piece of string that convert_to_utf8() cannot handle
   can be treated the same way.

 - handle_commit_msg() doesn't cope with a malformed line well, so
   die there for now.  We'll lift this even higher in later changes
   in this series.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Move the bulk of the code from builtin/mailinfo.c to mailinfo.c
so that new callers can start calling mailinfo() directly.

Note that a few calls to exit() and die() need to be cleaned up
for the API to be truly useful, which will come in later steps.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

When mailinfo() is eventually libified, the calling "git am" still
will have to write out the log message in the "msg" file for hooks
and other users of the information, but it does not have to reopen
and reread what it wrote earlier if the function kept it in a strbuf.

This also removes the need for seeking and truncating the output
file when we see a scissors mark in the input, which in turn allows
us to lose two callsites of die_errno().
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

There is a strange "if (!mi->cmitmsg) return 0" at the very beginning
of handle_commit_msg(), but the condition should never trigger, because:

 * The only place cmitmsg is set to NULL is after this function sees
   a patch break, closes the FILE * to write the commit log message
   and returns 1.  This function returns non-zero only from that
   codepath.

 * The caller of this function, upon seeing a non-zero return,
   increments filter_stage, starts treating the input as patch text
   and will never call handle_commit_msg() again.

Replace it with an assert(!mi->filter_stage) to ensure the above
observation will stay to be true.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NJunio C Hamano <gitster@pobox.com>

All callers of this function refrain from calling it when
mi->metainfo_charset is NULL; move the check to the callee,
as it already has a few conditions at its beginning to turn
it into a no-op.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

This requires us to pass the struct down to decode_header() and
convert_to_utf8() callchain.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NJunio C Hamano <gitster@pobox.com>

This requires us to pass the structure into check_header() codepath.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

This one is trivial thanks to previous steps that started passing
the structure throughout the input codepaths.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Earlier we got rid of two function-scope static variables that kept
track of the states of helper functions by making them extra arguments
that are passed throughout the callchain.  Now we have a convenient
place to store and pass them around in the form of "struct mailinfo",
change them into two fields in the struct.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

This requires us to pass "struct mailinfo" to more functions
throughout the codepath that read input lines.  Incidentally,
later steps are helped by this patch passing the struct to
more callchains.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

These two are the only easy ones that do not require passing the
structure around to deep corners of the callchain.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

In this first step, move only 'email' and 'name' fields in there and
remove the corresponding globals.  In subsequent patches, more
globals will be moved to this and the structure will be passed
around as a new parameter to more functions.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

With the previous steps, it becomes clear that the mailinfo()
function is the only one that wants the "line" to be directly
touchable.  Move it to the function scope of this function.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

With the previous two commits, we established that the local
variable "line" in handle_body() and handle_boundary() functions
always refer to the global "line" that is used as the common and
shared "current line from the input".  They are the only callers of
the last function that refers to the global line directly, i.e.
find_boundary().  Pass "line" as a parameter to this leaf function
to complete the clean-up.  Now the only function that directly refers
to the global "line" is the caller of handle_body() at the very
beginning of this whole callchain.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

This function has a single caller, and called with the global "line"
holding the multi-part boundary line the caller saw while processing
the e-mail body.  The function then goes into a loop to process each
line of the input, and fills the same global "line" variable from
the input as it needs to read more lines to process the multi-part
headers.

Let the caller explicitly pass a pointer to this global "line"
variable as an argument, and have the function itself use that
strbuf throughout, instead of referring to the global "line" itself.

There still is a helper function that this function calls that still
touches the global directly; it will be updated as the series progresses.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

This function has a single caller, and called with the global "line"
holding the first line of the e-mail body after the caller finished
processing the e-mail headers.  The function then goes into a loop
to process each line of the input, starting from what was given by
its caller, and fills the same global "line" variable from the input
as it needs to process more lines.

Let the caller explicitly pass a pointer to this global "line"
variable as an argument, and have the function itself use that
strbuf throughout, instead of referring to the global "line" itself.

There are helper functions that this function calls that still touch
the global directly; they will be updated as the series progresses.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Two helper functions use "static int" in their scope to keep track
of the state while repeatedly getting called once for each input
line.  Move these state variables to their ultimate caller and pass
down pointers to them along the callchain, as a small step in
preparation for making this entire callchain more reentrant.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NJunio C Hamano <gitster@pobox.com>

This way, we can lose a forward decl for decode_header().
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NJunio C Hamano <gitster@pobox.com>

This function wants to call find_boundary() and is called only from
one place without any recursing, so it becomes easier to read if it
appears after the called function.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Whether this loop is left via EOF/break or upon finding a
non-continuation line, the storage used for the contination line
handling is left behind.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

This does not make a difference within the context of "git mailinfo"
that runs once and exits, as flushing and closing would happen upon
process termination.  It however will matter when we eventually make
it callable as an API function.

Besides, cleaning after yourself once you are done is a good hygiene.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

We pre-increment the pointer that we will use to store something at,
so the pointer is already beyond the end of the array if it points
at content[MAX_BOUNDARIES].
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

In olden days we might have wanted to behave differently in
decode_header() if the header line was encoded with RFC2047, but we
apparently do not do so, hence this helper function can go, together
with its return value.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The called function checks if the second parameter is either a NULL
or an empty string at the very beginning and returns without doing
anything.  Remove the useless call.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NJunio C Hamano <gitster@pobox.com>