> On 23 Jul 2025, at 10:55 PM, Junio C Hamano <gitster@xxxxxxxxx> wrote:
>
> Aditya Garg <gargaditya08@xxxxxxxx> writes:
>
>> +imap.markAsRead::
>> + Choose whether to mark the sent message as read or not.
>
> Is this something user typically want to use a single setting,
> or would it often be per invocation? Especially with the new
> invoker in send-email, wouldn't it become more like "if I use
> imap-send to stuff things in my outgoing folder, they shouldn't be
> marked as read, but fcc copies send-email stuffs via imap-send
> should be marked as read" or something like that?
So whenever the user changes the folder, he can change this option too?
Also, we have a command line way as well.
>
>> + Default is `true`.
>
> If we never marked these messages as read, then this default is a
> breaking change, isn't it?
Curl always marked the messages as read, so how would it be a breaking change? As far as OpenSSL is concerned, it marks the messages as unread, but due to issues with the license, as long as people are compiling git themselves, I doubt people are using it.
Also, were people really using imap-send itself?
>
>> @@ -47,11 +47,12 @@
>>
>> static int verbosity;
>> static int list_folders;
>> +static int mark_seen = 1;
>
> IOW, this smells problematic.
>
>> + if (mark_seen) {
>> + ret = imap_exec_m(ctx, &cb, "APPEND \"%s%s\" (\\Seen) ", prefix, box);
>> + } else {
>> + ret = imap_exec_m(ctx, &cb, "APPEND \"%s%s\" ", prefix, box);
>> + }
>
> Why so many braces around single-statement blocks?
Will fix
>
>> @@ -1702,6 +1710,12 @@ static int curl_append_msgs_to_imap(struct imap_server_conf *server,
>> curl_easy_setopt(curl, CURLOPT_READFUNCTION, fread_buffer);
>> curl_easy_setopt(curl, CURLOPT_UPLOAD, 1L);
>>
>> + if (mark_seen) {
>> + curl_easy_setopt(curl, CURLOPT_UPLOAD_FLAGS, CURLULFLAG_SEEN);
>> + } else {
>> + curl_easy_setopt(curl, CURLOPT_UPLOAD_FLAGS, 0L);
>> + }
>> +
>
> Why so many braces around single-statement blocks?
>
> According to
>
> https://github.com/curl/curl/blob/master/docs/libcurl/symbols-in-versions
>
> CURLULFLAG_SEEN first appeared in 8.13.0; INSTALL says we require 7.61.0
> or later, so this may be OK.
>
